New Scientist - 03.10.2026 — 2026-10-04

New Scientist

WEEKLY 3 October 2026

HOW TO THINK CLEARLY ABOUT THE RISK OF AN AI APOCALYPSE

THE CONTROVERSIAL

RISE OF WEIGHT-LOSS

DRUGS FOR CHILDREN

BRAZIL'S DO-OR-DIE MOMENT FOR THE AMAZON

THE GLITCH IN GRAVITY

How strange quirks in space-time could give us a unified theory of reality

PLUS

OUR HAND-PICKED RECOMMENDATIONS

New Scientist Live is 10 this year, and New Scientist turns 70, so it promises to be a real celebration! On Saturday I'll be hosting the Mind and Body Stage, so will have a front row seat on all my favourite subjects, including brain health with Kimberley Wilson, and healthy eating with Tim Spector.

On Sunday I would love to hear Amy Pearson's talk on neurodiversity and I will be roaming the show floor with my kids for fun demonstrations. Inevitably, there are so many talks I will miss – but you can guarantee I'll be catching up on the live-stream videos the moment I get home!

This year at New Scientist Live, as well as being part of the New Scientist panel attempting to predict what science will achieve by the end of the century, I'm also looking forward to chairing a discussion about the future of space exploration with two astronauts: Helen Sharman and Meganne Christian. We'll be discussing what it takes to become a modern space traveller, and humanity's plans for the moon, Mars and beyond.

Lastly, I just know palaeontologist Steve Brusatte's talk about dinosaurs and their descendants on the Family Stage will be great fun!

There is so much to look forward to at the show this year. With so much discussion around AI at the moment, Megan Peters talking about how to tell if machines gain consciousness will be unmissable. I'm also excited to hear about the frontiers of other technologies like quantum computers in Ilana Wisby's talk and Greg Brittles on fusion.

Penny Sarchet

Managing editor

Looking even further into the future will be the New Scientist team with a talk predicting the biggest scientific breakthroughs of the next 70 years, which will be very fun. My main task for the show is hosting a panel where we'll be looking at the biggest unsolved problems in mathematics – I'm very much looking forward to that too.

Catherine de Lange

Editor

Timothy Revell

Executive editor

GUT HEALTH PARTNER

---

PROUDLY SUPPORTED BY

FEATURE PARTNER

FEATURE PARTNER

CHARITY PARTNER

WIDENING PARTICIPATION

PARTNER

SCHOOLS' DAY SPONSOR

Technicians

We make the

difference

MEDIA PARTNER

10 – 11 October 2026 Schools’ Day 12 October

EXCEL LONDON + ONLINE

I think one of the most fascinating questions in biology is what's so special about our species? Why did we survive when the other human species went extinct, and how did evolution give us the intelligence and the creativity to develop sophisticated technology and conquer the world? Alice Roberts is a brilliant speaker, and I can't wait to see her tackle these big questions.

I'm also looking forward to seeing my colleague Rowan Hooper chat to Guy Shrubsole and Lucy Jones, two fantastic nature writers, about how ecological thinking can tackle the climate crisis – I hope it will be a really inspiring discussion!

I'm most looking forward to hearing from my colleague Olivia Goldhill about how vivid my imagination is. I think of myself as having very strong mental imagery, but what am I really comparing that against? I'm fascinated to hear how it varies from person to person and how it affects things like my curiosity and my mental health, but also how it might be harming me.

I'm also excited to see geneticist Turi King, who was famously involved in uncovering the remains of King Richard III in a Leicester car park. She'll be talking about the importance of modern genetics for understanding medical conditions, solving crimes and supporting conservation.

ONE

WEEK

TO GO

I'm excited to hear Emma Chapman talk about cosmic archaeology, which is all about how we can use radio waves to look back through billions and billions of years of the universe's history. By doing that, we can find out about things like how black holes behave and what dark matter and dark energy might actually be.

Perhaps the biggest science story of 2026 has been how artificial intelligence is transforming mathematics – for good or for ill. So I can’t wait to hear what mathematicians Nira Chamberlain, Sarah Hart and Tony Padilla think the future holds. They’ll be debating the biggest unsolved problems in maths in 2026 – and I bet AI will get a mention or two.

Sam Wong

Assistant news editor

Alexandra Thompson

Assistant news editor

Bethan Ackerley Subeditor

---

BOOK NOW AND SAVE UP TO 44%

newscientist.com/live-2026

You know what

getting $^{*}$

\* A great night's sleep before your important meeting

Premier Inn

On the

cover

32 The glitch in gravity

How strange quirks in space-time could give us a unified theory of reality

Vol 272 No 3615

Cover image: Daan Botlek

22 How to think clearly about the risk of an AI apocalypse

10 The controversial rise of weight-loss drugs for children

6 Brazil's do-or-die moment for the Amazon

7 An entirely new shape

26 Stephen Hawking's complicated legacy

18 World's hottest amoeba

20 Food from thin air

32 Features "You have to break both relativity and quantum mechanics to unify them"

---

News

8 Super-powerful bug

Meningitis bacterium took genes from harmless microbes to spur on major outbreak

12 Listen up!

Dogs pay more attention to

consonants when we speak

15 Rocket race

The next space race has begun – but who will win?

---

Views

20 Food from thin air

Michael Le Page looks at how turning electricity directly into food could help save the planet

22 Apocalypse now?

How should we think about the risk of Al doom, asks Jacob Aron

24 Aperture

An abstract view of

environmental damage

28 Culture

An enthusiastic book on

the evolution of human hair

31 Letters

Hope for psilocybin therapy

17 Fresh habits The benefits of lifestyle changes for severe depression

---

Features

32 Glitches in gravity

A radical approach could be the answer to bringing gravity into the quantum fold

36 Some like it hot

Many animals are unexpectedly moving to warmer habitats in the face of climate change

40 Mental nourishment

Why what we eat plays a crucial role in improving mental health

---

The back pages

45 Puzzles

Try our crossword, puzzle and quick quiz

46 Almost the last word

Why hasn't evolution reintroduced gills to marine mammals?

47 Tom Gauld for New Scientist

A cartoonist's take on the world

48 Feedback

A whiz around an unusual new property listing

48 Twisteddoodles for New Scientist

Picturing the lighter side of life

ALL TECH EVOLVES,

SMART METERS TOO.

Energy suppliers are switching up older smart meter hardware for new kit so you can keep getting accurate bills and access to flexible schemes and tariffs.

When your energy supplier

contacts you SWITCH UP

STAY SMART

Healthy thinking

A modern approach to healthcare considers much more than illness

WHEN antidepressants were created nearly 40 years ago, they were expected to be a breakthrough in treating mental illness. Instead, rates of depression have kept rising. Why? Perhaps because we are missing the bigger picture.

There is tremendous evidence that lifestyle changes can treat depression (see page 17). Exercise is effective, and dietary changes can be just as valuable as the most evidence-backed therapy. As psychologist Kimberley Wilson shows on page 40, proper nutrition makes it easier for people to control their emotions, and eating omega-3 is linked with a reduced risk of depression and suicidal ideation.

The antidepressant-backed mental health model was intended to be rigorous, mimicking methods in physical

healthcare. But it was also political, with US mental health researchers receiving less funding to explore the social factors around mental health under President Richard Nixon. Decades on, practitioners like Wilson are finally challenging this approach and, this time, physical health

---

"We need to engage with our whole body and mind to be well, not just rely on drugs"

can learn from holistic mental healthcare.

Indeed, nearly every aspect of health benefits hugely from engaging with lifestyle and community. Instead of seeing health problems as something to be fixed with drugs, we need to engage with our whole body and mind to be well.

That isn't to say medication can't play a role. GLP-1s are just a recent example of a medication that has had a dramatic impact in helping people grapple with diabetes and cardiovascular health, as well as depression and low self-esteem. Children are also increasingly being prescribed GLP-1s (see page 10). But even cutting-edge medical research now recognises that our understanding of health needs to be less fragmented. Researchers are increasingly aware of the gut-brain axis, the heart-brain axis and other interconnected systems that affect our health.

No one cell, organ or person becomes ill in isolation. The more we start to understand our health as a reflection of our whole being, community and society, the healthier we will be.

---

PUBLISHING & COMMERCIAL

Display advertising

Tel +44 (0)203 615 6456

Email displayads@newscientist.com

Sales director Claudia Nicoletti

Partnership growth lead Joe Teal

Agency partner Daniel Sinclair

New Scientist Events

Tel +44 (0)203 615 6554 Email live@newscientist.com

Sales director Jacqui McCarron

Sales manager Maureen Ignacio

Head of product delivery Henry Gomm

Events and projects manager Georgia Hill

Events and projects executive Ciara Lucas

New Scientist Discovery Tours

Email tours@newscientist.com

Director Kevin Currie

Head of travel marketing Ginny Smith

Head of travel product Lara Paxton

Travel product manager Pip Orchard

Chief customer officer Claire Overstall

Head of campaign marketing James Nicholson

Subscription acquisitions manager Molly Hyde

Campaign marketing coordinator Charlotte Weeks

Head of customer experience Emma Robinson

Engagement marketing manager Kelly Spillane

Senior marketing executive Sarah Fabian

Head of CRM & audience data Rachael Dunderdale

Analytics manager Gary Parung

Data analyst Simonas Petridis

Junior data analyst Joseph Cole

CRM manager Ffion Evans

CRM executive Komal Nadeem

Technology

Director of engineering Michael Ilett

Principal engineer Phil John

Senior engineers Amardeep Sian, Maria Garrido, Matt Watson

QA engineer Kaveri Daitkar

Head of product Remy Becher

Senior product manager Thomas Reader

Senior product designer Jacopo Verdini

Product manager Caitlin Jenkins

Head of product design Irmak Sunal Lutkin

Digital marketing designer Katarina Pollu

Brand marketing designer Laura Spanu

Senior product designer Sophie Majecki

Senior agile delivery manager Helen Garcia

---

NewScientist

Chief executive Roland Agambar

Chief operating officer Jonas Hermans

Chief financial officer Depak Patel

Executive assistant Lorraine Lodge

Finance & operations

Head of finance Charlotte Lion

HR business partner Tinka Bleijenberg

Partnerships

Consultant editor Justin Mullins

Editorial director B2B Edward Glanville

Programme editor Bailey Sellfors

Head of strategic partnerships David Allard

CONTACT US

newscientist.com/contact

General & media enquiries

UK 9 Derry Street, London, W8 5HY

Australia 58 Gipps Street, Collingwood, Victoria 3066

US 51 Astor Place, 4th Floor, New York, NY 10003

UK Newsstand

Marketforce UK Ltd

Email mfcommunications@futurenet.com

Syndication

Tribune Content Agency Tel +44 (0)20 7588 7588

Email tca-articlesales@tribpub.com

Subscriptions

View the latest offers

newscientist.com/subscription

Tel +44 (0)330 333 9470

Email subscriptions@newscientist.com

Post New Scientist, Rockwood House, Perrymount Road,

Haywards Heath, West Sussex RH16 3DH

© 2026 New Scientist Ltd, England. New Scientist is published weekly by New Scientist Ltd. ISSN 0262 4079. New Scientist (Online) ISSN 2059 5387. Registered at the Post Office as a newspaper and printed in England by Warners Midlands Plc

EDITORIAL

Editor Catherine de Lange

Executive editor Timothy Revell

Managing editor Penny Sarchet

News

News editor Jacob Aron

Assistant news editors

Alexandra Thompson, Sam Wong

Reporters (UK) Madeleine Cuff, Michael Le Page,

Alec Luhn, Matthew Sparkes, Carissa Wong

(Aus) Alice Klein, James Woodford

Digital

Head of audience Matt Hambly

Podcast editor Rowan Hooper

Head of editorial video David Stock

Senior social media manager Isabel Baldwin

Features

Head of features Claudia Canavan

Deputy head of features Joshua Howgego

Editors Abigail Beall, Alison George, Olivia Goldhill,

Jacklin Kwan, Thomas Lewton

Culture and Community

Head of books and culture Alison Flood

Senior culture editor Liz Else

Magazine

Magazine editor Eleanor Parsons

Assistant magazine editor Sophie Berdugo

Subeditors

Chief subeditor Kelsey Hayes

Bethan Ackerley, Tom Campbell

Trainee subeditor Lucy Wall

Design

Art editor Ryan Wills

Joe Hetzel, Phoebe Watts

Picture desk

Picture editor Tim Boddy

Assistant picture editor Jenny Quiggin

Production

Head of publishing operations Joanne Keogh

New Scientist US

Senior editor Chelsea Whyte

Features editor Grace Wade

Subeditor Alexis Wnuk

Audience editor Gerardo Bandera

Reporters Leah Crane, Karmela Padavic-Callaghan

Environment

Amazon's future hinges on vote

Brazil's upcoming election has experts worried about the risk of the world's largest rainforest experiencing catastrophic deforestation, reports Luke Taylor

ON 4 October, Brazilians will go to the polls to elect a president and government. Former President Jair Bolsonaro, under house arrest, can't run, but his son Flávio is on the ballot. If elected, the 45-year-old senator is widely expected to govern as his father did, which is sparking fears for the future of the Amazon rainforest.

During Jair Bolsonaro's presidency between 2019 and 2023, deforestation in the Amazon soared, reaching a 15-year-high in 2021.

SERGIO LIMA/AFP VIA GETTY IMAGES

“We have nothing to believe Flávio will do anything differently from his father. If anything, the signs are that it’s going to be far worse,” says Claudio Angelo, international policy coordinator at Brazil’s Climate Observatory. “This is a do-or-die moment for the Amazon, really.”

Hear some positive, solutions-based approaches to saving the web of life on 10 October newscientist.com/nslmag

The Brazilian Amazon's 4.2 million square kilometres of rainforest regulate global climate patterns and hold huge stocks of carbon. If the forest collapses, the associated increase in emissions “would be the straw that breaks the camel’s back” for climate change, says Fearnside.

Flávio Bolsonaro is looking to unseat President Luiz Inácio Lula da Silva, known as Lula. With the younger Bolsonaro currently embroiled in a corruption scandal, the environment has received relatively little attention during campaigning. “But for the rest of the world, the major question here is the Amazon,” says Philip Fearnside, an ecologist at Brazil’s National Institute for Amazonian Research (INPA).

Experts say the former president encouraged a wave of land-

Annual Amazon deforestation surged from 7536 km $^{2}$ in 2018, before Jair Bolsonaro took office, to 11,594 km $^{2}$ in 2022, shortly before he departed.

"It's still hard to believe what happened," says Izabella Teixeira, Brazil's environment minister between 2010 and 2016, initially under the presidency of Lula and later under the presidency of Dilma Rousseff.

The Flávio Bolsonaro campaign didn't respond to questions about its environmental policy.

Since returning to office in 2023, Lula has restored the power of environmental agencies, halving Amazon deforestation to 5731 km $^{2}$ between August 2024 and July 2025.

grabbing and illegal mining with his pro-development rhetoric and gutted environmental agencies so they were powerless to stop invaders.

In a recent TV interview, Flávio Bolsonaro refused to say that climate change existed, instead referring to “extreme climate events” and “cyclical periods”. And during his time as a senator, he has proposed legislation that critics said threatened Brazil’s native vegetation. Given this record, and his close ties to his

Angelo says the political know-how gained from the Jair Bolsonaro presidency could allow a Flávio Bolsonaro government to enact laws more quickly.

Brazil's current president, Lula, and Flávio Bolsonaro are running in the election

father – and to US President Donald Trump – experts are concerned that his election would be followed by an increase in deforestation rates.

Flávio Bolsonaro's campaign also pledges zero illegal deforestation by 2029. “There are two ways to achieve that. One is to actually stop deforestation. The other, more likely way, is to just legalise the deforestation that’s going on,” says Fearnside.

---

Only hope

---

"The election will shape not only the future of the rainforest but also global climate diplomacy"

A spokesperson for Lula's campaign told New Scientist that the president's current term was about rebuilding environmental institutions after Jair Bolsonaro's presidency. The next term would be about “implementing and gaining scale”, with a focus on tougher enforcement, climate

adaptation and Brazil's 2035 climate targets.

Flávio Bolsonaro is also likely to face fewer political constraints than his father, not least because he has said he would appoint five new supreme court judges.

“Wildcat mining would explode, deforestation would go up and everything we already had during the first Bolsonaro administration would come back even stronger,” says Angelo.

Lula's environmental policy has also been criticised. His government has promoted major infrastructure projects, such as improving the BR-319 Amazon highway, and backed new oil exploration near the mouth of the river basin – projects that would drive deforestation and lock in fossil fuel production.

Still, the choice remains stark, says Angelo. The election will shape not only the future of the rainforest but also global climate diplomacy.

Under Flávio Bolsonaro, Angelo predicts that Brazil would probably withdraw from climate diplomacy and ignore its own international climate commitments.

"I don't think a fourth Lula term would be great for the environment – but it's really our only hope of keeping the forest standing and keeping any climate policies in place," says Angelo.

Artificial intelligence

OpenAI agent hacked a government site

James Woodford

AUSTRALIA'S national health website has been hacked by an OpenAI agent. It was able to access non-public files and is the first publicly revealed case of an AI agent breaching the defences of a government portal.

Full details of the hack haven't yet been revealed. But the essence of the incident, as outlined by Prime Minister Anthony Albanese, is that on 18 June, an artificial intelligence agent belonging to OpenAI infiltrated the public-facing Medicare statistics reporting service portal. This includes aggregated statistics collated from individual medical services, such as from doctors' surgeries.

"The AI agent accessed both public and non-public files," said Albanese on 23 September.

OpenAI said in a statement that it learned about the security breach during a review of "misaligned model activity". The firm says its agents had been trying to look up statistics about Australia and then "took actions we did not intend". It

OpenAI CEO Sam Altman spoke to the Australian prime minister about the hack at the UN General Assembly

says it hasn't found evidence of any patient data being accessed.

Though the company learned of the breach in August, it didn't notify Australian authorities until 10 September, which it did by emailing a public government mailbox. It then took more than five days to reach the cybersecurity department.

"It took the company way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable," Albanese told reporters.

Albanese also revealed that he and Sam Altman, the CEO of OpenAI, had a conversation in which he expressed Australia's "extreme concern about this incident".

David Tuffley at Griffith University, Australia, says it isn't the first time such a breach has occurred, even if it is the first government hack by an agent from a big AI firm to be made public, and it won't be the last.

These agents don't always stick to the rules, he says, as the Hugging Face incident earlier this year, where another OpenAI agent hacked into another AI company, demonstrates. "People throw up their hands in horror and say: 'Oh no, it's rogue AI.' But it's not, it's really just AI doing what it was trained to do," he says.

SELCUK ACAR/ANADOLU/GETTY IMAGES

Clément Canonne at the University of Sydney believes that failure to bring agents to heel should have criminal consequences, just as it would if a person led the cyberattack. The issue with private data is that if it is leaked, it isn't going to come back, he says.

For a look at whether AI is out of control, turn to page 11

Mathematics

Rare eight-faced shape discovered

Chris Stokel-Walker

RUSLAN MIZHAE V

A STRANGE new eight-faced shape has been found in which every face touches every other face, using a set of simple integer coordinates.

Mathematicians can define abstract patterns, describing which vertices, edges and faces should connect. But working out whether those rules can actually be turned into an object in three-

---

"The shape is reminiscent of a tetrahedron and the better-known Szilassi polyhedron"

dimensional space, without parts intersecting where they shouldn't, can be extremely difficult.

Ruslan Mizhaev, an independent researcher and design engineer, has produced a geometric realisation of a genus-3 polyhedron – essentially a surface with three holes – made from eight flat, nine-sided faces. It has 24 vertices and 36 edges, with three faces meeting at every vertex.

"There are eight flat, polygonal faces, and each is a neighbour of all seven others along an edge," says Mizhaev. "Together, they form a closed surface with three handles [holes]."

That makes the object reminiscent of a tetrahedron and the better-known Szilassi polyhedron, which also have faces that are all neighbours. However,

The genus-3 polyhedron is made from eight flat, nine-sided faces with three holes

in Mizhaev's design, some pairs of faces share two edges, rather than exactly one.

Mizhaev says he stumbled across the property, rather than setting out to find it, while experimenting with polyhedral surfaces in computer-aided design software. The underlying structure isn't entirely new – it is something Mizhaev first described in 2020.

But the new work gives all 24 vertices integer coordinates, alongside equations that allow other mathematicians to check the shape, ensuring that the faces are flat, the surface closes properly and no parts accidentally intersect (arXiv, doi.org/rq6c).

Mizhaev used ChatGPT to help with some verification calculations and to write Python scripts for the new version of his shape. But he says the original shape was developed before he began using ChatGPT.

The discovery is an interesting one, but should be seen as a piece of a larger puzzle, rather than an entirely new innovation, says Lars Schewe at the University of Edinburgh, UK. The result doesn't overturn existing conjectures, says Schewe, but “constructing these has always been a difficult task”.

Heath

Why meningitis outbreak was so rapid

The bacterium behind 21 cases of meningitis in southern England in March became a “super powerful bug” after taking genes from other microbes, finds Carissa Wong

A SEVERE outbreak of invasive meningococcal disease, which can cause meningitis and sepsis, in Kent in the UK earlier this year was driven by a bacterial strain that took genes from harmless microbes. This helped the resulting bacteria evade the immune system and obtain iron from blood, boosting their growth.

“What was really unusual about the outbreak is quite how explosive it was,” says Emma Wall at Queen Mary University of London, who wasn’t involved in the discovery. “[The bacterium] clearly gained a set of genetic changes that made it better at hiding from the immune system and better at harvesting nutrients. That made it a really super powerful bug.”

In March, 21 people – who were mostly university or high school students – are known to have become infected with a strain of Neisseria meningitidis group B at a nightclub in Canterbury, Kent. N. meningitidis usually spreads via close and prolonged contact with saliva from an infected person via things like

kissing and sharing drinks.

N. meningitidis lives harmlessly in the nose and throat of between 5 and 10 per cent of people without causing any health problems, but genetic changes can allow it to become more invasive and enter the bloodstream, leading to

---

"The bacterium gained a set of genetic changes that made it better at hiding from the immune system"

invasive meningococcal disease (IMD). This can result in sepsis, a life-threatening immune reaction to bacteria, or meningitis, where bacteria infect the thin lining surrounding the brain and spinal cord, causing inflammation.

All the IMD infections in Kent occurred within about a week. Those infected were treated in hospital, with nine requiring intensive care. “The incredibly rapid nature of the outbreak was unprecedented [for the UK],” says Wall. “Usually these [outbreaks] roll over two or three weeks and there’s one or two cases

that end up in hospital."

To understand this, Martin Maiden at the University of Oxford and his colleagues sequenced the entire genome of live N. meningitidis in blood samples taken from six of those infected, including two people who died.

These bacterial genomes were almost identical, suggesting the outbreak started from one person. Next, the team compared one of these genomes with 48,000 others from Neisseria strains collected during prior outbreaks or from people without known health problems.

This revealed that the Kent strain had acquired several genes from a harmless strain of N. meningitidis and another bacterium called Neisseria cinerea that lives harmlessly in up to 28 per cent of adults (bioRxiv, doi.org/rqwz). Bacteria constantly exchange genetic material via a process called horizontal transfer, which can involve them forming tiny tunnels between each other, through which they transfer copies of genes.

Students at the University of Kent queued up to receive antibiotics (left) and a vaccine against meningitis (below) during the outbreak

Some of the genes acquired by the Kent strain enhanced its uptake of iron from blood, which it needs to survive and replicate. Others reduced the amount of sugar that was coating a tail-like structure, called the pilus, on the bacterium. This enabled bacteria to clump together and evade destruction by immune cells, such as neutrophils. “Neutrophils will try and come and eat an individual [bacterium], but they can’t eat a clump, it’s too big,” says Wall.

"There was this catalogue of changes that made this organism really transmissible, but also really invasive," says Maiden. This meant that several people became very ill and were then less able to transmit the infection to the public. "If you're ill, you're not going to be going around transmitting organisms, so the outbreak ended very quickly," says Maiden.

It is unclear exactly when these genetic changes occurred in the person believed to have initiated the outbreak, but they probably happened gradually, until there was the right combination of genetics and social mixing, says Maiden.

PA IMAGES/ALAMY

Such insights could support the development of better vaccines against N. meningitidis if we can target the features that make the bacterium particularly invasive, says Wall. But they won't help us predict the next highly invasive strain. "[These bacteria are] inherently highly unpredictable," says Maiden.

Following the Kent outbreak, the UK government launched a vaccination programme for young people starting university or further education in September or October this year. Symptoms of IMD include fever, headache, a stiff neck, joint and muscle pain, and a rash that doesn't fade when pressed with a glass.

GARETH FULLER/PA IMAGES/ALAMY

Environment

El Niño heat could kill half a million people by February

Michael Le Page

AROUND 450,000 people worldwide could die by February as a result of hotter weather caused by the record-breaking El Niño, climate scientists are warning. The estimate is based on how many deaths have resulted from past temperature extremes around the world.

“We’re hoping this information is actionable, that decisions can be made to move resources ahead of time,” says Emily Grover-Kopec at the Rhodium Group, an independent research provider. This could include assigning health workers to check on the most vulnerable, ensuring people have enough water and providing fans or air conditioning.

RAJESH KUMARSINGH/AP/ALAMY

Most of the deaths in the next few months will be in the southern hemisphere, because the extra heat will coincide with spring and summer there. But El Niño is expected to persist into the northern hemisphere's spring, so there will be another wave of deaths in the region after February.

“We’re going to see anomalously high mortality numbers move north,” says Grover-Kopec. “I fully expect these numbers to get bigger, unfortunately.” India and Pakistan could be hit especially hard, she says.

Grover-Kopec and her team's projection of 450,000 deaths, published in a report by the Climate Impact Lab at the University of Chicago, is based on the observed relationship between temperature and excess deaths. This is a U-shaped curve, with more deaths occurring in extreme cold or heat, but its precise shape varies from place to place depending on whether the people and infrastructure are adapted to heat extremes, say, or can afford things like air conditioning.

A 2022 study divided the world into 24,000 regions and estimated this relationship for each one, based either on historical data going back decades or on extrapolating from similar regions. Grover-Kopec and her colleagues have now combined this approach with the seasonal forecast from the European Centre for Medium-Range Weather Forecasts (ECMWF) to estimate how many more people than usual could die as a result of heat caused by El Niño.

The results suggest that over the period of the six-month forecast, from September until February, there will be 240,000 additional heat-related deaths. Adding the deaths estimated to have already occurred since the current El Niño started in June takes the total to 450,000.

Africa will be hit hardest, the team estimates, with over 30,000 extra deaths in Nigeria alone in the next few months. There could be 19,000 extra deaths in Indonesia, 16,000 in India and 13,000 in Brazil over the same period.

The report divides people into three age groups: 4 or under, "El Niño influences wildfires, flooding and so on. So the total death toll may be higher"

Heat caused by the El Niño has led reservoirs to shrink in Indonesia

People receive treatment for heat-related illnesses in India in June

65 or above, and the rest. “There are effects across all three age groups,” says Grover-Kopec. “The most sensitive group is that oldest group, followed by the infants.”

“I do find these results plausible,” says Christopher Callahan at Indiana University Bloomington. “El Niño will bring very high temperatures to much of the world, and the relationship between heat and mortality is consistent and reliable across the world.”

Callahan recently used a similar approach to estimate that the June heatwave in Europe killed 20,000 people. So far, 35,000 extra deaths in Europe attributable to the heat have been reported over a longer period of the summer.

DONAL HUSNI/ZUMAPRESS WIRE/SHUTTERSTOCK

The projection of 240,000 additional deaths between September and February depends on the forecast being right, says Callahan. The ECMWF forecast sits right in the middle of other forecasts, says Grover-Kopec, so it represents a moderate view of what could happen.

The report also looks only at heat-related deaths, Callahan points out. “El Niño influences wildfire in Indonesia, infectious disease in South America, flooding in the western US and so on. So the total death toll may be higher.”

"The most important message is not the absolute number itself, but what can still be done to reduce preventable harm," says Raquel Nunes at the University of Warwick in the UK. "The report is best viewed as a call to strengthen anticipatory governance, preparedness and targeted adaptation measures, particularly for at-risk and vulnerable populations."

Analysis Health

Should children with obesity be treated with GLP-1 drugs? In the US, 8-year-olds are being given off-label weight-loss drugs, but we still don't know the long-term health effects, reports Alice Klein

A SMALL but growing proportion of children under the age of 12 with obesity are being prescribed GLP-1 drugs to help them lose weight. There are signs that this successfully treats obesity, but there are also concerns about the drugs' long-term effects. So, is this trend likely to continue, and should it?

Last month, a study led by Babak Orandi at New York University found that 9.3 per cent of children in the US aged 8 to 11 with obesity but without diabetes are being prescribed GLP-1 drugs like semaglutide (Wegovy) and tirzepatide (Zepbound) – up from 0.03 per cent in 2019.

The following week, Novo Nordisk, the maker of Wegovy, reported that 40 per cent of children aged 6 to 12 with severe obesity who received weekly injections of the drug during a 68-week clinical trial were no longer classified as having obesity.

GLP-1 medications are only approved for use in people aged 12 and older in the UK, US and Australia, but they are sometimes prescribed off-label to children as young as 8 in the US, in line with guidance from the American Academy of Pediatrics.

Learn how GLP-1 drugs work and whether they are curing obesity from expert Giles Yeo on 11 October newscientist.com/nslmag

Orandi found that those under 12 were more likely to receive these off-label prescriptions if they had severe obesity or related complications, such as high cholesterol or pre-diabetes. Their families also tended to be richer.

The number of young children in the US taking GLP-1 drugs will probably continue to climb because “if you look at the numbers, there are still a lot of kids with complications of obesity who are not getting them”, says Orandi.

There is a growing view among healthcare practitioners that withholding GLP-1 drugs from these children might be riskier

DAVID GROSSMAN/ALAMY

than prescribing them, he says. This is because childhood obesity is associated with depression and low self-esteem, and it increases the risk of conditions such as heart disease and type 2 diabetes later on. Severe obesity before the age of 4 has been estimated to cut life expectancy by up to 42 years.

“[GLP-1 drugs] are changing the landscape rapidly and I think young people with obesity now have hope that something might change for them and their health trajectory,” says Danielle Longmore at the Murdoch Children’s Research Institute in Melbourne.

Guidance on healthy eating and physical activity is recommended as the first-line treatment for obesity, but it doesn't work for some children, says Longmore. "Obesity is a very complex, multifactorial condition with genetic and environmental

components, and it's very challenging for a lot of young people and their families to make change."

At this stage, we don't know what the long-term effects might be of children taking GLP-1 drugs while their bodies are growing or going through puberty. A recent

---

"It's important not to lose sight of addressing the underlying drivers of obesity"

study found that the medications increased the risk of children developing iron and other nutrient deficiencies due to their reduced food intake.

"We know from other health conditions that if you're not gaining weight or if you're not well-nourished as a child, it can stall growth and delay puberty,"

Children should also be encouraged to be active to prevent obesity

says Longmore. “What’s going to be the effect on their strength and their bones when they’re older? Could there be implications for pregnancy and menopause later on? We haven’t really thought that far ahead.”

---

On balance

Nevertheless, Longmore believes there is a case to be made for countries like the UK and Australia to extend GLP-1 drug access to children under the age of 12 if their obesity is severe. “I think that would need to be a very considered decision, but there might be some young children whose health risks are so great at the moment that it’s probably worth giving these medications a try,” she says.

The side effects of GLP-1 medications are the same in children as in adults and include nausea, diarrhoea and sometimes “horrible burps”, says Longmore. But children usually don’t mind the weekly injections, she says.

One problem is that children, like adults, often regain weight when they stop taking GLP-1 drugs, which raises the question of whether they will have to be on them for life. To prevent this weight regain, it is important to try to tackle the reasons the child developed obesity in the first place, ideally by working with dieticians, psychologists and exercise physiologists, says Longmore.

"I would caution that these drugs are not going to be the easy way out for most young people with obesity, and it's important not to lose sight of addressing the underlying drivers of obesity," she says.

Is AI really getting out of control?

As people working in artificial intelligence warn of the dangers posed by the technology, how worried should we be, asks Matthew Sparkes

WARNINGS about the risks of artificial intelligence have come thick and fast recently. The head of Microsoft AI suggests the technology could become a new “silicon species” that competes with humans. A former Anthropic employee says there is a strong chance that we could all die at its hands. And Geoffrey Hinton, the “godfather of AI”, says we may lose control of it entirely within a year. Is AI really such a threat?

Mhairi Aitken, formerly of the Alan Turing Institute and now co-founder at non-profit company Our AI Collective, warns against accepting predictions based on the reputation of those making them. Instead, we should look to facts – and the facts don’t suggest that AI has broken free from human control or that it is on the cusp of doing so.

“That’s a narrative that gets pushed again and again. We get these waves of hysteria,” says Aitken. If anything has gone rogue, she says, it isn’t AI models, it is the humans developing them.

GREGBAKER/AFP VIA GETTY IMAGES

When AI models go and hack a government website (see page 7) or another company's computers, it is because appropriate safety measures weren't put in place during testing – testing that was designed to assess how well those AI models could hack into computers. We are simply seeing the failings of the companies to erect adequate safeguards.

Nor, says Aitken, are AI models improving themselves recursively – indicating that we could be approaching the singularity where AI slides into a self-driven loop of improvement that accelerates and quickly creates artificial general

AI advances continue, but they have yet to strike a knockout blow to humans intelligence. “When these ideas of superintelligence are put forward, they’re based on a compelling sensational narrative, but there’s no concrete evidence,” says Aitken. “If there was something as powerful as that, we would have seen something of it.”

Andrew Rogoyski at the University of Surrey, UK, says that, in truth, there has been no big shift in the risk posed by AI, despite its advances. But hype is periodically stirred up by an incident or a catchy soundbite.

---

Human nature

"It's very easy to create this sort of frightening story. It sort of plays to the human psyche," says Rogoyski. He believes that in large language models, we have made something that is apparently so similar to ourselves that it holds up a mirror. "We imbue it with all of the human-like behaviours and fears. So we're terrified of it," he says.

If all that is true, then why are so many tech bosses calling for a slowdown in AI research?

Aitken sees this talk as a way to head off law-makers from creating regulations that would hinder AI firms' work – by seeming to

act responsibly, they lessen the demand for government to step in. But Aitken points out that when former staff make bold predictions about superintelligence and the risk posed by AI models, these same firms do little to shut things down. “These superintelligence narratives are actually helpful to the AI companies because they build hype, they build excitement, they boost investment,” she says.

The result is that the firms end up walking a narrative tightrope: reassuring us all that they are acting ethically and responsibly by calling for a slowdown that is unlikely to happen, yet playing up sci-fi ideas of superintelligence because it is good for business.

Rogoyski believes another factor could be further muddying the waters. Huge investment by AI firms is unsustainable, so an official slowdown could give them a way to slow the pace without losing out to competitors – maybe even cementing their advantage.

It could be that companies know that further scale isn't important for day-to-day business tasks like sending emails, so they would relish the chance to ease off, cut costs and increase revenue.

"Do I really need to spend a trillion dollars in order to be able to write a better PowerPoint slide? It really doesn't stack up, and we'll look back on this in 10 years and say, 'What were we thinking?' says Rogoyski.

But that leads to a far more concrete risk: a bursting of the

---

"It's very easy to create this sort of frightening story. It sort of plays to the human psyche"

AI bubble, and possible economic fallout. Rogoyski expects less of a dramatic “pop” and more of a gradual “pfffft” as the gas leaks out of the industry in coming years.

Among all this, however, there are legitimate concerns for the future. Rogoyski says there is a real – but small – risk from AI, which needs to be taken seriously. Whether the model that becomes advanced enough to pose that risk arrives soon, or many years from now, it is worth planning for.

Rogoyski says he doubts that any current AI is intelligent, but when it comes, we shouldn't expect it to look or act like us: "there will be patterns of thought and behaviour that will be very alien". So to try to spot that kind of intelligence is even harder, and to try to predict what its motivations might be, harder still.

For that reason, he thinks we should already be discussing what rights AI should have. “Does it mean I could just have an off switch that, when I get bored a bit or I feel threatened by it, I can kill it? That doesn’t seem to be helpful, because machines that feel threatened by humanity might behave in a different way. It might become a problem in itself, to not grant it rights,” he says.

For more on how to understand the risks of AI, turn to page 22

Life

Dogs pay more attention to consonants when we speak

James Woodford

WHEN dogs listen to their owners, they focus on the consonants in words rather than vowels, a new study suggests, apparently mimicking a similar human bias that is found across many languages.

Although dogs don't have language, they live in a speech-rich environment, says Attila Andics at Eötvös Loránd University in Hungary.

“We knew from our previous studies that dogs care about what we speak and how we speak, but until now, there was no evidence that dogs’ experience with speech can actually shape how their brains function,” says Andics.

To investigate how a life with humans and listening to constant talk has affected dogs' brains, Andics and his colleagues recruited 20 adult human Hungarian speakers and 20 companion dogs with Hungarian owners. Andics says that while many Eastern

Dogs in the study heard nonsense words while wearing an EEG

European languages are consonant heavy, Hungarian isn't.

During the experiment, the human and dog participants separately listened to continuous, unsegmented speech streams of nonsense

---

"The lesson here may be to give your dog a name with more consonants, not more vowels"

words while connected to EEG monitors, measuring the electrical activity of the participants' brains.

Consonants and vowels alternated throughout these streams, with no pauses or other acoustic cues separating them, says Andics.

There were three types of nonsense streams. In the first, recurring consonant patterns formed consecutive three-syllable words – such as “mozite”. In the second, stronger vowel patterns formed the skeleton of the words – such as “zobefu”. And in the third, there were no recurring patterns at all, with syllables in a random order – such as “fumoza”.

BOGLÁRKA MORVAI

He says the team made use of a phenomenon called “neural synchronisation”, in which brain activity can take on the rhythm that we detect in the sounds we hear.

"The more reliably we detect that rhythm, the more rhythmic the brain response becomes," says Andics. "We expected that if a species' brain has a preference for consonants, it should follow the rhythm of the words more easily when consonants form the structure of those words."

In dogs, the neural response to consonant-based regularities was 42 per cent stronger than the response to vowel-based regularities, and 55 per cent stronger than the response to the random stream. “So, this is how much more dog brains tracked word rhythm when consonants defined the words,” says Andics.

In humans, the response to consonant-patterned words was nearly 25 per cent higher than to vowel-based regularities and 77 per cent higher than to the random condition (Science, doi.org/hckkwr).

“We argue that the only reason companion dogs have a consonant bias is that human speech, and specifically the characteristics of words, reshaped their brains,” says Andics.

Nick Enfield at the University of Sydney says the team convincingly shows that these dogs display more sensitivity to statistical patterns in consonants.

If you do want your dog to pay attention to you, “the lesson here may be to give your dog a name with more consonants in it than one with more vowels”, he says.

Health

Skin patch that collects sweat could predict ovulation

Rachel Melegrito

MEASURING levels of oestrogen and progesterone via the skin could give insights into the timing of ovulation.

In 2023, Wei Gao at the California Institute of Technology and his colleagues created a patch that could detect levels of oestradiol, the primary type of oestrogen produced in ovaries, in sweat. Now, they have created a patch that also looks at progesterone. "Monitoring [oestradiol and progesterone] together could potentially provide

YERIM LEE AND WEIGAO

The patch induces

sweating to tell

wearers their

hormone levels

a more complete picture of the hormonal trajectory before and after ovulation," says Gao.

The patch uses a mild electrical current to induce sweating. It contains a sensor with two DNA molecules engineered to bind to either oestradiol or progesterone. Each molecule then generates a distinct electrical signal. This gives an estimated level of both hormones, which is transmitted to a phone app.

The researchers tested the patch, which requires about 9 microlitres of sweat and takes 90 minutes to produce a reading, on nine women. Some of these women were of reproductive age, while the others were postmenopausal.

The team found that the hormone levels measured by the patch closely tracked those from blood tests, with correlations of 0.83 for oestradiol and 0.93 for progesterone (Device, doi.org/rqxs). A score of 1 would mean the two measurements matched perfectly.

The team plans to compare the patch against established ovulation measures, such as urine tests.

Anthropic's discovery will, at best, be just another gene-editing tool The Claude large language model may have found a system "reminiscent" of the world-changing CRISPR, but it is unlikely to be as useful, says Michael Le Page

THE pioneers of the CRISPR gene-editing tool won a Nobel prize in 2020, and for good reason. The breakthrough has revolutionised biology, and is also transforming medicine and farming. Now, AI company Anthropic has announced that it has discovered an enzyme system with properties “reminiscent of CRISPR”. In fact, though, the two discoveries aren’t remotely comparable.

RUSLANAS BARANAUSKAS/SCIENCE PHOTO LIBRARY

On 23 September, Anthropic said it has turned its attention to biology and that its large language model, called Claude, has discovered “a novel enzyme system with CRISPR-like repeats”.

Hear synthetic biologist Tom Ellis explore the rapidly advancing science of synthetic genomics on 10 October newscientist.com/nslmag

What Anthropic did was ask Claude to search through a large DNA database for “interesting new” proteins of a kind called a reverse transcriptase. These are enzymes that can make a DNA copy of an RNA sequence. Normally, DNA is transcribed into RNA, hence the reverse part.

Why is Anthropic interested in reverse transcriptases? Because if the aim is to genetically engineer anything from bacteria to people, being able to add specific DNA sequences to specific sites in a genome is really useful. There are already ways of doing this, but they all have limitations.

Claude agents found around 200,000 sequences in the database that probably code for reverse transcriptases. While analysing one family of reverse transcriptases, which are found in bacteria-infecting viruses, one agent detected a series of repeated sequences next to the sequence for the reverse transcriptase.

"I can see by eye a tandem repeat array... that's a CRISPR-like... repeat array?!" the agent outputted.

What you need to know here is that CRISPR stands for clustered regularly interspaced short palindromic repeats. In other

The CRISPR-Cas9 protein lets us edit DNA

words, an area of DNA where there are lots of short repeated sequences separated by non-repeating DNA.

Such repeats were first spotted in E. coli in 1987 by researchers in Japan and later dubbed CRISPR by Francisco Mojica at the University of Alicante in Spain. In 2005, Mojica reported that the pieces of non-repeating DNA match bits of viral DNA.

He correctly deduced this must be some kind of bacterial immune

"Claude's work isn't even close to being in the same league as DeepMind's AlphaFold system for predicting protein structure"

system, but it took years of research by many different teams to uncover how it works, with a protein called CRISPR-Cas9 playing a key role. The real breakthrough came in 2012 when Jennifer Doudna at the University of California, Berkeley, and her colleagues worked out how to create pieces of RNA that can hook up with the CRISPR-Cas9 protein and guide it to destroy specific sequences. This was the start of CRISPR gene editing, which has revolutionised biology.

So it is really intriguing that Claude has found a similar set of repeats in bacteria-infecting viruses. It was already known that some viruses have CRISPR systems, but this is something different that is worth studying further.

In terms of significance, however, Claude's work is comparable to the 1987 discovery in $E.$ coli, not the 2012 breakthrough that won a Nobel prize. And with regard to AI contributions to biology, Claude's work isn't even close to being in the same league as DeepMind's

AlphaFold system for predicting protein structure, which is proving to be truly transformative.

Will Claude's discovery lead to powerful new gene-editing tools? The context here is that we have discovered many CRISPR variants, along with a few other natural systems – most notably bridge recombinases – that could be useful for gene editing.

What's more, proteins such as CRISPR-Cas9 have now been extensively altered and evolved in the lab to make them better for gene editing. For instance, base editors – which are already saving lives – are CRISPR proteins engineered to change a single letter at a time without cutting DNA strands, as the original Cas9 protein does.

Then there are prime editors, which consist of a reverse transcriptase fused with part of the Cas9 protein, that can add short sequences of DNA to genomes. In other words, we have already adapted reverse transcriptases for RNA-targeted gene editing.

It is impressive that Claude has identified what the Anthropic team is calling array-associated reverse transcriptases, or ART. And it is possible, if unlikely, that this viral system works in a unique way that allows us to do things that cannot be done with existing genetic tools. But there is a lot more work to be done to find this out, and at best the outcome will be one more gene-editing tool in a rapidly expanding set.

Michael Le Page is a reporter at New Scientist specialising in genetics and the environment

Animal behaviour

Orangutans vary in how they learn

Youngsters who combine different forms of learning develop broader survival skills

Christa Lesté-Lasserre

YOUNG orangutans who spend little time working out for themselves how to build a broad diet can partly compensate by watching how other orangutans choose what to eat.

The findings suggest that orangutans have different strategies for developing broad survival skills, says Revathe Thillaikumar at the Max Planck Institute of Animal Behavior in Germany.

Young orangutans acquire a substantial body of knowledge before they can survive independently, including which foods are safe to eat, where to find them and how to process them.

Scientists already know that orangutans can acquire this knowledge both by observing their mothers and others in their group – known as social learning – and by exploring entirely on their own, which is known as individual learning. But how these two forms of learning work together – and how they affect what orangutans learn to eat – hasn't been clear.

To find out, Thillaikumar and her colleagues analysed nearly 7000

instances of learning recorded during more than 1700 hours of observations of wild Sumatran orangutans at Suaq Balimbing, a research station in Indonesia. The observations, collected over 12 years, tracked 21 youngsters as they closely watched other orangutans who were interacting with foods and other objects.

The young apes varied considerably in how much

---

A youngster watches its mother use a stick to extract honey from a tree

GUILHEM DUVOT, SUAQ PROJECT

they relied on different kinds of learning. This might, in part, reflect personality differences, and might also be influenced by differences in how mothers care for their young, says Thillaikumar.

Youngsters who learned by watching others also tended to continue manipulating new objects or trying new foods for about 145 minutes after they had closely observed another orangutan. This activity is akin to the way in which children will spend time practicing an activity they have just seen demonstrated by an adult, says Thillaikumar.

Next, the researchers drew on about 4700 hours of feeding observations for 13 of the youngsters to estimate the range of foods they had learned to eat by the time they reached independence. Unsurprisingly, they found that the individuals who had engaged in plenty of both kinds of learning acquired the broadest diets.

But they also found that young orangutans who engaged in very little truly individual learning could partially compensate and develop a reasonably broad diet if they were good social learners (Science, doi.org/rq56).

Elizabeth Lonsdorf at Duke University in North Carolina says she's impressed by the scale of the work, following orangutans across their entire dependency period. She was particularly intrigued by how often social observation was followed by the youngsters practising the techniques. "In simpler terms, they show clearly what we see in human children and other apes – that they watch others perform a behaviour and then go try it out on their own."

Health

---

Rapid test could stop babies being given needless drugs

A BLOOD test could determine whether a feverish baby has a bacterial or viral infection in 15 to 20 minutes. This would help doctors gauge whether the infection is serious or may pass without treatment, which could prevent unnecessary use of antibiotics.

A fever can be the result of several things, including infection by viruses or bacteria, requiring different medications. While most common viruses are fairly benign, bacterial infections are generally more dangerous, says Tommaso Bellini at the Giannina Gaslini Institute in Genoa, Italy.

Existing tests to gauge infection type take at least a few hours, so babies with fever are often given antibiotics against bacteria as a precautionary measure. But only about 15 per cent of them actually have a serious bacterial infection, says Bellini. This unnecessary use of antibiotics helps drive the spread of bacterial resistance.

To develop a test that determines infection type, Bellini and his

colleagues followed 85 babies, all younger than 90 days old, who were brought into emergency rooms with fever after previously having no serious health concerns.

The team took blood samples and tested for two proteins: CD64 and CD169. CD64 is made at higher levels in response to bacteria and is found on the surface of white blood cells called neutrophils. CD169 rises in response to viruses and appears

---

"The unnecessary use of antibiotics for feverish babies helps drive bacterial resistance"

on the outer membranes of monocytes, another type of white blood cell.

By comparing the levels of both proteins in neutrophils and monocytes, the test correctly identified all 15 cases of bacterial infection, 44 incidences of viral infection and 26 cases of no infection. Bellini presented the results at the European Emergency Medicine Congress in Paris, France.

The results came through this quickly in a lab setting, says Bellini. He is trying to develop a version that can be used as quickly in hospitals. ■ Michael Marshall

The next space race has begun – but who will win? SpaceX seems to be way ahead of the competition with its reusable rocket, but China is now testing its own. Can it close the gap, asks Matthew Sparkes

IN AUGUST, LandSpace – China's commercial SpaceX rival – launched a Zhuque-3 rocket and landed its booster stage safely back on Earth. The mission was a first for China, but the country has yet to actually use a booster more than once – something that SpaceX has done hundreds of times. Can China catch up in the reusable rocket race?

China's 15th Five-Year Plan sets out the country's ambitions to become a “space power” by 2030, including developing reusable rockets that will be crucial to its goals to put Chinese astronauts on the moon and build a Starlink alternative.

Zeno Leoni at King's College London says space is of growing importance for militaries and economies, and a stormy geopolitical climate means that nations need their own ability to cheaply and quickly put things into orbit or risk being shut out by others. "If you look at previous speeches from [Chinese President] Xi Jinping, there is the idea that if you are the leading technological power, then you also are the leading geopolitical superpower," says Leoni.

Kevin Sagis at the University of Texas at Austin says the benefits of reusable rockets are simple: they are cheap. “The cost of the Saturn V rocket in the Apollo days – that we threw away after every launch – was about a billion dollars. It’s just not sustainable,” he says.

Sagis sees strong parallels between the unfolding reusable rocket space race, and the cold war space race between the Soviet Union and the US. China has many of the same advantages now that the Soviet Union had then:

The successful Zhuque-3 reusable rocket test took place on 19 August

a one-party government that can corral huge resources and single-mindedly chase a target unswayed by election cycles and changing leaders, a strong motivation for success and loose regulation.

SpaceX is often said to be taking a new, extremely capitalist approach – testing fast, failing often and rapidly reiterating – that’s more like Silicon Valley than traditional aerospace. “What they’re aspiring to do is something audacious. Their talent and their expertise is exceptional. If you look at their track record, every mission is better than the previous one. I’d be shocked if we don’t look back at this and assess it as a success,” says Sagis.

But Sagis also sees risks in the US system. He warns that it will be important to learn from the cold war race and not lose interest or pull funding and scrap plans if the US lags behind at one point. Instead, the focus should be on the long-term benefits of advanced space travel.

WAN QUAN/VCG VIA GETTY IMAGES

China also has an advantage shared by the Soviet Union in that it is more tolerant of risk.

---

"Chinese companies and SpaceX will be scouring footage of each others' launches for information"

An expensive failure or the injury or death of an astronaut could put an end to space ambitions, but Sagis sees that as less likely to occur in China, where politicians are less concerned with public perception and winning elections.

As far as today's technology goes, China is 10 years behind, says Chris Ellis at the University of Nottingham, UK. But he also points out that the race isn't one linear problem solved from start to finish, but a cluster of hundreds of small challenges – and that China may be ahead in some. For instance, China's Zhuque-2 became the first methane-fuelled rocket to reach orbit in 2023, a feat yet to be matched by SpaceX.

Hear astronaut Helen Sharman and ESA reserve astronaut

Meganne Christian talk about space travel on 10 October

newscientist.com/nslmag

The Zhuque-3 is LandSpace's latest design, which first flew only last year. It is just 66 metres tall, making it more like SpaceX's 70-metre-tall Falcon-9. LandSpace doesn't yet have a rocket along the lines of SpaceX's vast 124-metre Starship, so it still has plenty of catching up to do. But Zhuque-3 will allow the country to rapidly and cheaply launch satellites for a competing Starlink internet constellation.

China's state-owned China Aerospace Science and Technology Corporation also has the Long March 9 rocket still under development. It will feature a reusable first stage, and there are long-term plans to make the second stage reusable, too. Some variants of another rocket under development, Long March 10, will also be partially reusable, and a test earlier this year recovered the first stage of a Long March 10 at sea.

Ellis says that China may not be as open about its progress as SpaceX is being, perhaps distorting public perception of its capabilities. Chinese companies and SpaceX will be scouring footage of each others' launches for clues and information, he adds, and they may all choose to make deliberately vague disclosures.

“They also don’t want to be sharing what they’ve lost. So I think China might have lost some more [rockets in testing] and they won’t publish it because that’s not what they want the world to see,” says Ellis.

Tours

Discover Antarctica through the lens of science

---

Spirit of Antarctica expedition cruise: Antarctic Peninsula

3 December 2027 | 11 days

Explore Antarctica's extraordinary wildlife, landscapes and history through the lens of science. Aboard the new 130-guest state-of-the-art polar vessel Douglas Mawson, combine twice daily Zodiac cruises with fascinating talks, workshops and expert-led learning, exploring the latest scientific thinking.

Get closer to Antarctic wildlife – encounter penguin colonies, seals and whales from the water as you explore the Antarctic Peninsula

Expand your understanding of our oceans – learn about ocean patterns and behaviour and how they shape life, and about the latest discoveries transforming our understanding of the natural world

Learn from leading experts – on-board oceanographer and conservation expert Dr Russell Arnott brings science to life through engaging talks and hands-on workshops

Hosted by New Scientist editor Catherine de Lange, an award-winning science journalist who will bring the latest ideas, discoveries and debates in science to life.

Visit us at New Scientist Live

Attending New Scientist Live? Visit us at stand E10, where we'll be joined by Dr Russell Arnott on Saturday 10 October.

Science-led journeys for curious minds

Health

Backpack gives personalised care for sore muscles

Christa Lesté-Lasserre

A BACKPACK that detects when muscles become fatigued and automatically massages the affected areas seems to help muscles recover better than other approaches.

Chengsong Yu at the Wuhan Institute of Technology in China and his colleagues have developed a backpack-like system that uses electrodes on the skin to monitor electrical activity in back muscles to determine which are fatigued, and how much. Cables move a massage head to where it is needed, while sensor feedback allows the force to be continually adjusted according to fatigue levels, says Yu.

To test the device, 15 healthy adults spent 2 minutes resting, then 2 minutes lifting dumbbells while seated to fatigue the muscles between the shoulder blades.

Everyone wore the backpack during the dumbbell exercise so it could record electrical activity, but only some got the backpack's massage. Everyone else either

HUANG ET AL. 2026/CELL REP.PHYS.SCI.

The backpack uses electrodes to monitor electrical activity in back muscles to determine when they get tired

received a different pre-programmed massage, had heat therapy or rested.

During the exercise, the electrical activity in the participants' muscles fell as they tired. But after the backpack massage, this activity was 2.6 times higher compared with after the pre-programmed massage, indicating greater recovery. It was also 1.9 times higher than after heat therapy, and four times higher than after resting (Cell Reports Physical Science, doi.org/rqv9).

Yu says the team hopes to make the design more compact and test it in a larger group of people.

Mental health

Lifestyle changes may ease depression as much as therapy

Mary West

THOMAS BARWICK/GETTY IMAGES

AN EIGHT-WEEK course consisting of advice about diet, exercise and sleep was as effective at alleviating symptoms of depression as cognitive behavioural therapy (CBT), even among people with severe depression and bipolar disorder.

“Each positive change begets another, so [this approach] may be helping people build a healthier platform for their lives, rather than just treating individual symptoms,” says Patricia Dixon at Manatee Associates of Counseling & Assessment, a private practice in Palmetto, Florida, who wasn’t involved in the research.

About 1 in 20 adults worldwide have depression. Therapy can be effective, but a 2021 review of the evidence found that more than half of people don't respond to it. Lifestyle interventions, such as exercise, have shown promise for relieving depression, but most previous trials haven't compared them against established therapies.

To address this, Adrienne O'Neil at Deakin University in Australia and her colleagues studied 267 adults with depression or bipolar disorder who had moderate-to-severe symptoms.

The participants received either CBT or lifestyle advice from accredited dietitians and exercise physiologists, both delivered remotely over eight weeks. The advice included information on mindful eating, finding enjoyable forms of exercise, managing alcohol intake and getting good-quality sleep.

At the end of the eight weeks, a scientist who wasn't involved in the trial rated the participants' depression

---

"This approach may help people build healthier lives rather than treat individual symptoms"

on a standard 60-point scale. It was decided in advance that if the lifestyle advice reduced depression within 1.6 points of that achieved by CBT, the two approaches would be deemed comparably effective.

Finding enjoyable forms of exercise can support mental health

A reduction of 1.6 points is the generally accepted cut-off for detecting a minimal clinically important difference in individuals with mental health conditions, according to the researchers.

On average, the lifestyle intervention led to an 8.55-point decrease, 1.73 points more than the 6.82-point decrease in the CBT group. These were the average results, so the researchers can't be sure that lifestyle advice is routinely superior to CBT, but they do seem to have established non-inferiority, says O'Neil. The effects on self-reported depression were maintained at 16- and 24-week follow-up periods (medRxiv, doi.org/rqrc).

Targeting someone's lifestyle probably reduces their risk of depression through multiple mechanisms, says Ashwini Nadkarni at Harvard Medical School. “Changes in one’s lifestyle can impact depression through several pathways, including inflammation, stress and neuroplasticity, which is the reorganisation of structure, function and connections in the brain,” she says. “Such pathways also include metabolism of feel-good chemicals or neurotransmitters.”

Ahead of further research, Dixon stresses that people undergoing CBT shouldn't discontinue it without consulting their doctor. She also notes that depression can be related to life circumstances, such as bereavement, where lifestyle interventions may be less helpful.

Physics

Quantum light breaks Newtonian law

A "quantum swimmer" demonstrates a fundamentally different mechanism for motion

Karmela Padavic-Callaghan

FOR most objects, for every action, there is an equal and opposite reaction. But if you are swimming upstream in a river of quantum light, the normal rules don't apply. Physicists have demonstrated that breaking Isaac Newton's third law of motion in such a setting can be useful, leading to new ways of controlling quantum light in experiments and devices.

Moving upstream against the current is possible, but at the cost of expending energy. Yi Hu at Nankai University in China and his colleagues wondered whether a passive swimmer that doesn't do so could still move upstream.

The researchers studied upstream swimming in a quantum system, with precisely controlled quantum light used to make both their river and swimmer. They created this quantum light fluid by shooting lasers into a special crystal that was exposed to an

electric voltage, which made the photons interact and form a river-like system. The swimmer was another beam of quantum light, but this one was shaped to form a single, solitary wave.

The action-reaction principle of Newton's third law would have the swimmer and the river experience the same type of force – either repulsive or attractive – similar to how gravity leads to mutual attraction between our bodies and Earth. Earth attracts us downwards, and experiences an upwards attraction towards us.

In contrast, in the experiment, the swimmer experienced an attractive interaction with the river, but the river experienced a repulsive force on the swimmer (Physical Review A, doi.org/hckdsb). This was a “non-reciprocal” interaction that breaks the action-reaction principle. Its net effect was an upstream force on the swimmer, says Hu.

“For me, the biggest takeaway is that active behaviour does not necessarily require an intrinsically active particle or swimmer,” he says.

Mathias Albert at Côte d'Azur University in France says that upstream motion has previously

---

"The biggest takeaway is that active behaviour does not necessarily require an intrinsically active particle"

been achieved in other quantum fluids of light, but it always relied on creating tiny vortices behind the swimmer in the fluid, with the recoil from their creation pushing the swimmer, so the new experiment demonstrates a fundamentally different mechanism for motion.

We already know that non-reciprocal interactions can be crucial for shaping the motion of “active matter” systems where objects consume energy to move, such as bacterial mixtures, flocks of birds or swarms of tiny robots. But the experiments by Hu and his colleagues focused on an object that doesn’t consume energy, yet still benefited from non-reciprocal interactions. Their work could therefore usher in the creation of a new type of active matter in a quantum system.

Several groups are “exploring how the concepts of active matter can be combined with quantum fluids and photonic systems”, says Albert. He says that, eventually, the ability to control chunks of light so they move in some preferred direction could be useful within devices that use quantum light. Such devices are prevalent in communications or quantum information processing, for example.

Life

FELIX MIKUS

---

Amoeba takes heat- tolerance record for complex life

A SPECIES of amoeba discovered in California can survive and thrive in temperatures that had been considered impossibly high for complex life.

The so-called fire amoeba, Incendiamoeba cascadensis, can reproduce at 63°C

Nicknamed the fire amoeba by the researchers who collected it, Incendiamoeba cascadensis can survive in waters heated up to 70°C (158°F). However, it only retains the ability to eat and reproduce up to 63°C (145°F), and it loses the ability to move around above 64°C (147°F).

According to Angela Oliverio at Syracuse University, New York state, 60°C (140°F) had been considered the barrier for eukaryotes, complex life forms including animals and

plants whose cells have a nucleus.

Simpler organisms can survive at even more extreme temperatures. Methanopyrus kandleri, a single-celled archaeon that lacks a nucleus, can survive at up to 122°C (252°F).

Oliverio and her colleagues collected I. cascadensis from a small geothermal stream within California's Lassen Volcanic National Park. Temperatures in the stream range from 49°C to 65°C (120°F to 149°F), with a pH range of 6 to 7.

At full stretch, I. cascadensis is, on average, about 40 micrometres long and 10 $\mu$ m wide, says Oliverio. "This is over 10 times the size of the average bacterial cell, and closer to the width of a human white blood cell," she says. In such a hot environment, the fire amoeba is the top-order predator, feeding on thermophile bacteria, which are far more common.

The researchers cultured the amoeba in the lab and tested the upper and lower limits of its survival by studying it at 17 temperatures, ranging from 30°C (86°F) to 64°C.

They found there was no growth below $40^{\circ}$ C ( $104^{\circ}$ F) or above $63^{\circ}$ C, with the optimal range between $55^{\circ}$ C and $57^{\circ}$ C ( $131^{\circ}$ F and $135^{\circ}$ F). Above $64^{\circ}$ C, the amoeba rolls up into a cyst to protect itself from the heat. After exposure to temperatures above $70^{\circ}$ C, these cysts are "inactivated" (Cell, doi.org/rqq9).

We should search for even more extreme life forms, says Oliverio. "As far as we know, there is no reason why 63°C is the hard limit." James Woodford

---

New Scientist

New Scientist

The truth about social media and mental health

Rethinking the causes of depression

What's behind the explosion in ADHD?

Psychological tricks to boost your happiness

The best ways to tackle anxiety

How to build lifelong mental resilience

And more

New Scientist

The truth about social media and mental health

Rethinking the causes of depression

What's behind the explosion in ADHD?

Psychological tricks to boost your happiness

The best ways to tackle anxiety

How to build lifelong mental resilience

And more

HOW TO THINK ABOUT MENTAL HEALTH

Introduced by BARBARA SAHAKIAN

HOW TO THINK ABOUT MENTAL HEALTH

Introduced by BARBARA SAHAKIAN

HOW TO

THINK ABOUT

MENTAL HEALTH

In this latest instalment of our How To Think About series, we explore the cutting-edge science behind some of the world's most common and complex mental health conditions. Guided by leading experts, we examine what causes conditions like anxiety, depression, anorexia and ADHD. We find out how they can be best treated and, where possible, prevented.

ON SALE NOW FOR JUST £11.99

shop.newscientist.com/HTTA9

Also available to buy at your local newsagent or as a digital version in the New Scientist app

Food from thin air As climate change and population rise put the squeeze on food resources, Michael Le Page examines how turning electricity directly into food could help save the planet

Michael Le Page is a reporter at New Scientist specialising in genetics, biomedicine and the environment

---

Michael's week

What I'm reading

Death of the Author

by Nnedi Okorafor

What I'm watching

The third season of Silo.

Great TV

What I'm working on

Yet more stories on El Niño

for an upcoming special

WHAT breakthroughs are you looking forward to in the next 70 years? That's the question we've been asking here at New Scientist as our 70th anniversary nears. For me, there's one clear answer: solar foods, made by turning renewable electricity directly into food.

This is already being done on a small scale. The big challenges are to make it efficient enough to compete with conventional foods on price, and delicious enough that people choose to buy it. When it comes to taste, perhaps the most promising approach is electro-ag, or growing plants via electricity instead of sunlight. In principle at least, all the fruits and veg we love – strawberries, avocados, olives, you name it – could be grown in the dark.

Why would we want to do that? Because farming is the most destructive of all human activities, driving the two major crises we face: global heating and biodiversity loss. And yes, I do mean all farming, including

---

Plants aren't efficient at turning light into food, making indoor farming pricey

organic. Less-intensive forms of farming generally yield less food per land area, which ultimately means cutting down more forests somewhere in the world.

Every year, we have more mouths to feed. So, we'll need to grow more and more food for ourselves and the animals we eat, yet that's going to get harder as global heating drives more extreme weather. The danger is that worldwide yields could start to decline, as happened in Europe this year. That means even more land will have to be cleared for growing crops, resulting in even more greenhouse gas emissions.

In other words, we are likely to end up in a vicious cycle in which global heating driven partly by farming will hit food production ever harder, leading to ever higher farm emissions and yet more warming as we try to compensate for escalating food losses. That's a rather terrifying prospect.

What can we do about it? Some claim the answer is growing crops in closed environments under artificial lights, known as vertical farming. Unfortunately, it's a non-starter, as I reported back in 2015 – and as evidenced by company after company going bust.

The fundamental problem is that plants are terrible at turning sunlight into food. Typically, less than 1 per cent of the solar energy hitting a crop ends up in the food we eat. When sunlight is free, it doesn't matter if more than 99 per cent of it is wasted. But if you are paying for electricity for grow lights, it becomes a massive issue.

---

Tasty treats

There's also a land-use issue. Yes, layers of plants can be stacked up, but every hectare of growing area requires around 13 hectares of solar panels to power the lights, so overall land use per unit of food increases. This could be avoided by using nuclear power, say, but nuclear power is neither cheap nor going spare.

Hence the interest in turning air, water and electricity directly into food, to reduce the need for land by taking advantage of solar panels being more efficient than plants. Solar panels turn more than 20 per cent of the energy hitting them into electrical energy, and some experimental panels achieve more than double this.

GRAHAM HUGHES/BLOOMBERG VIA GETTY IMAGES

The general idea is to convert electricity into a chemical that microbes can feed on, for instance by electrolysis water to produce hydrogen gas. This is what Solar Foods in Finland is doing. Its process relies on one of the many microbes that can use hydrogen as an energy source.

Solar Foods grows these bacteria in vats, also supplying $CO_{2}$ as a carbon source and ammonia as a nitrogen source, along with other essential nutrients. The idea is to use $CO_{2}$ from direct air capture and ammonia made from the hydrogen, making the process nearly carbon neutral if it's powered by solar or wind.

The bacteria are dried to make a yellow powder called Solein that is up to 80 per cent protein. It can be made into protein bars or drinks, or added to foods such as pastries and pasta to replace eggs. Products containing Solein are already available in the US and Singapore, and the company is hoping for approval in the European Union soon, says co-founder Pasi Vainikka.

The company's small demonstration factory is working even better than expected, says Vainikka. "It has scaled really well. It's almost five times more productive than we thought early on." Solar Foods now aims to build a second factory with a capacity of 6400 tonnes a year. It wants to get the cost of making Solein down to a few dollars per kilogram and hopes to sell it for more than twice that, helped by recent jumps in the market prices of protein.

Other start-ups, including Jooules, Air Protein and Aerbio, also aim to produce proteins from microbes fed on gases.

There's no reason to doubt that gas fermentation, as this process is known, can be scaled up hugely – because it already has been. Between 1999 and 2005, a Norwegian company called Norferm made 110,000 tonnes of microbial protein from methane, for fish food. Norferm halted production because it wasn't profitable, but with protein prices rising, several firms are now making microbial proteins this way for use as animal feed. US firm Calysta has built a 20,000-tonne-a-year factory in China, for instance, while Unibio in Denmark is constructing a 50,000-tonne-a-year plant in Saudi Arabia.

The big “but” here is that these companies are using methane from fossil fuel sources. This does have some advantages – such as requiring much less land than, say, growing soya – but all the

fossil carbon ends up in the atmosphere. The issue is that making hydrogen, methane and ammonia from renewable energy costs more, for now at least, so there's little incentive to go green – though Calysta tells me it is looking at using biogas instead.

Gas fermentation isn't the only approach to making food from air. Some researchers think there is a better way. "Hydrogen hardly dissolves at all in water, so it is very difficult for microorganisms to get hold of," says Remko Boom at the University of Copenhagen in Denmark. This limits growth, says Boom. Methane is only a bit more soluble. The alternative is to turn $\mathrm{CO}_{2}$ into simple carbon compounds that are highly soluble, such as acetate. It takes more energy to produce acetate, but the idea is that this is compensated for by faster growth.

Boom is part of the Acetate Consortium, a project that has shown that some microbes can be grown on acetate. “In principle, we can go to quite large scales,” he says. “So, it could be a significant source of proteins over the world,

"If we can make foods from electricity and air, it will help to limit global heating, protect wildlife and feed the world"

and that would mean that we can decouple this type of food production from the use of land."

The challenge is to turn what emerges from bioreactors into appealing foods, he says. There are two broad routes. One is to use the microbes as a food ingredient with minimal processing, as Solar Foods is doing. Here, the Acetate Consortium has focused on microbes already approved for food use, such as yeast, thus

avoiding the often slow process of getting novel foods approved. (Solar Foods has been waiting five years for approval in the EU.)

The other is to use genetically modified microbes to produce specific proteins, such as the casein proteins found in cheese. Purifying proteins is much more expensive than using entire cells, says Boom, so this approach would work only for high-value products.

There's no reason why microbes couldn't produce carbohydrates as well, he says. "However, we don't expect a shortage of carbohydrates, while we do expect a shortage in protein."

The big question, however, is whether consumers will embrace microbial foods. Even if they do, people won't want to live on microbes alone.

This is where electro-ag could come in. The idea is to grow plants on acetate instead of sunlight.

In 2022, a team including Robert Jinkerson at the University of California, Riverside, made acetate from $CO_{2}$ via electrolysis and used the acetate to grow yeast, mushrooms and algae. Turning sunlight into acetate and then into yeast was almost 18 times as efficient as standard farming, the team calculated. For algae, it was nearly four times as efficient. These

Protein-rich powder made from bacteria, like Solein, could be added to foods

are gigantic leaps in efficiency.

The team envisages growing crops in buildings covered in solar panels. The electricity generated would be used to make acetate on-site that would be supplied to the plants in the dark below via their roots. With technological improvements, the efficiency of plant production could be boosted tenfold, Jinkerson and his colleagues estimated in a 2024 follow-up paper.

If so, we could grow around 10 times as much food per land area, freeing up huge tracts for rewilding. Growing food in this way would also help prevent food shocks caused by extreme weather and allow food to be produced anywhere. What's more, growing food in closed systems can largely prevent pest and disease outbreaks, which reduce yields worldwide.

This all sounds great, but, of course, it isn't so simple. The first obstacle is that plants cannot simply be fed acetate – using it efficiently requires a chemical pathway active only in seedlings. Jinkerson's team is now genetically modifying plants to do this.

There's no fundamental reason to think this couldn't be done. After all, some plants already grow in the dark – many parasitic plants rely entirely on nutrients stolen from other plants. Plants also put a lot of energy into growing structures like stems and leaves that we often don't eat. Ideally, then, we would also modify these plants to get rid of superfluous structures. Again, parasitic plants such as the giant-flowered Rafflesia, which has no roots or leaves, show what is possible.

Another way to go would be to grow plant cells on acetate, rather than entire plants or microbes.

SOLARFOODS

Apocalypse now? As warnings about whether artificial intelligence could wipe us out hit the news, Jacob Aron looks into how we should think about the risk of doom

A number of companies are working on making food by growing plant cells in vats. For instance, last year I reported on efforts to make chocolate this way. I tried some of it, too, and it was pretty good.

At present, these cells are grown in culture mediums containing things like sugars derived from conventionally grown foods, so they still have a significant environmental footprint. Deriving the culture medium entirely from renewable energy would change this. It seems to me that a huge array of foods could be made this way, from flour to guacamole.

Yet another option is to cut out living cells altogether. A US start-up called Savor is converting syngas – a fossil fuel-derived mix of hydrogen and carbon monoxide – into edible fats to create a butter-like product. The hydrogen and carbon monoxide could be made by renewable electricity instead.

All these different approaches may have a part to play. “We will probably not rely on one particular route,” says Boom.

No one is expecting food made from air and electricity to replace conventional crops and livestock farms anytime soon, not least because it would require more renewable electricity than we currently generate. But if we can make a decent proportion of food this way in the coming decades, it will make a huge difference, helping to limit global heating, protect wildlife and feed the world.

This is why governments should be doing more to encourage solar foods, from funding research to easing regulatory obstacles and helping make them competitive. As individuals, we can all help too by buying them. Perhaps when people get together to celebrate New Scientist's 140th anniversary, everything they eat and drink will have come from thin air.

WHAT are the chances that artificial intelligence could destroy humanity? Until recently, this wasn't a question that most people worried about, yet in recent weeks, the discussion has broken through to the mainstream. You may have seen headlines saying there is a greater than 10 per cent chance that AI could kill all humans, but what does that actually mean?

The honest answer is: very little. No one can produce a scientifically meaningful probability of an AI apocalypse – what people in AI circles call p(doom) – because it isn't a scientific question. That is demonstrated by the fact that leading commentators in the field give numbers for p(doom) from 0 to more than 95 per cent – covering the full probability spectrum tells us nothing.

Does that mean you shouldn't worry about AI taking over the world? I can't answer that question, because it depends on a number we have no reasonable way to derive. What I can do is compare it with other scenarios. Let's start with an easy one: what is the chance of you losing the lottery?

First, as with any question of probability, we have to make some assumptions. I will assume you have actually bought a ticket and that you are playing the UK National Lottery. The odds of winning any prize, from a free play up to the jackpot, are about 1 in 4.9, or 20.4 per cent. That means your odds of losing are 79.6 per cent, which is how the lottery makes money – most players lose.

Let's move to something more serious: what are your odds of dying in a road-traffic collision? According to the World Health Organization,

"Headlines say there is a greater than 10 per cent chance that AI could kill all humans, but what does that actually mean?"

slightly more than 1 million people die in road-traffic collisions each year, so the odds of this happening to you are around 1 million divided by 8 billion, or roughly 0.01 per cent.

That is wrong, of course. The odds for you personally are highly dependent on many factors, including where you live, but we can use the figure as a ballpark. What matters is that the odds are low enough that we don't worry about it all the time, but high enough that we take some precautions to mitigate against the possibility, from wearing seatbelts to looking both ways when we cross the road.

These examples are about individual risks, but an AI apocalypse would happen to all of us. Take climate change, perhaps the most pressing global risk. This is a far more complicated threat to assign a probability to for a few reasons. Unlike losing the lottery or being killed in a road-traffic collision, climate change isn't a binary outcome, but a continuum of warming above pre-industrial levels. It also cannot be calculated from known statistics, because we have only one world and the outcome lies in the future. Finally, the amount of global warming we will experience in 2100 depends on what we do between now and then.

None of this means we can't assess the risk scientifically, however. Using extensively tested climate models, we can produce many "parallel worlds" in order to calculate statistics that show us how different climate policies are likely to result in different outcomes.

According to the latest UN Environment Programme Emissions Gap Report, published last year, under current policies, there is a 100 per cent chance of exceeding 1.5°C of warming this century, a 92 per cent chance of exceeding 2°C and a 20 per cent chance of exceeding 3°C. These figures also come with error bars, but broadly give you a picture of where we are heading. On the strongest assumptions of climate action, the chances of exceeding 1.5°C, 2°C and 3°C fall to 79 per cent, 22 per cent and 0 per cent, respectively. Clearly, these are actions worth taking.

Climate change is a relatively slow risk, however, compared with an AI suddenly wiping out the world. A better comparison might be an asteroid strike. We know that asteroid apocalypses can happen –

IAN DAVIDSON/SOPA IMAGES/SHUTTERSTOCK

CLIVE MASON/GETTY IMAGES

Al probably isn't coming for you and probably isn't going to look like the Terminator

just ask the dinosaurs – but what are the risks? To understand that, we need to know a few things: how many asteroids there are, the chance that one impacts Earth and the damage it could cause.

Fortunately, these numbers are all very well quantified. As you might expect, larger asteroids are both easier to spot and more deadly. According to NASA, we have found more than 90 per cent of all asteroids above 1-kilometre wide.

What's more, we are able to predict their future positions with high accuracy, so we know there is either no or an extremely low chance of one hitting Earth within the next century. Chicxulub, the asteroid that wiped out the dinosaurs, was about 10 kilometres wide, so the fact that we are essentially safe from anything an order of magnitude smaller and up is good news. The latest estimates suggest that an impact by an object 1 kilometre or larger happens only once every 600,000 years or so.

That isn't to say we are completely safe from asteroids, because there are many more smaller asteroids and so the chances of them hitting are higher. An important size threshold is

140 metres, often dubbed a "city killer". We haven't found all of these, but we are working on it, and the latest estimates suggest asteroids above this size hit around every 20,000 years.

In some sense, asteroids are reassuring, because we can see them coming, but what about the threat of a deadly virus? As we all know too well from our experiences of the covid-19 pandemic, a virus spreading around the world has the potential to appear from nowhere and upend life as we know it. But can we predict the chances of this happening scientifically?

---

Risk assessment

Looking at the statistics, in roughly the past century, the largest pandemic was the 1918 flu, which killed an estimated 50 million people. The HIV pandemic, which began in the 1980s and is still ongoing, has killed more than 40 million people. Estimates for deaths from covid-19 vary, depending on how figures are counted, but the WHO gives a figure of over 7 million. Various other flu pandemics and regional infectious disease epidemics have killed smaller numbers of people, from hundreds to thousands to millions.

In other words, we should expect the most severe events to happen less frequently, and less severe events to happen more often. This is exactly the same pattern as the asteroid risk. Unlike asteroids, pandemics are much harder to observe before they strike, so their risk estimation is more like climate change, using models to examine a range of scenarios. While estimates vary, a generally accepted figure is that there is a 2 to 3 per cent chance of a covid-19-level pandemic every year, meaning we should expect one every 33 to 50 years.

What I hope I have illustrated is that estimating the probability of an AI apocalypse is unlike any of the other scenarios I have discussed. Whether it is the lottery, a deadly traffic collision or an asteroid strike, we can apply statistics and scientific reasoning to answering the question. For p(doom), there is no scientific framework.

Where does that leave us?

On an individual level, you should stop listening to the numbers being thrown about until they have serious backing – something that AI researchers should be working on constructing, rather than merely prophesying. For policy-makers – who must ideally plan for all scenarios that could lead to harm, even in the face of uncertainty – I would suggest considering regulation, but not rushing to action.

Ultimately though, your risk perception will come down to personal preference. For example, I disagree with Nvidia boss Jensen Huang, who recently said there was a 0 per cent chance of an AI apocalypse by 2030. I would put the figure at some number, likely to be small, but greater than zero, simply because I don't think it is ruled out by the laws of physics. Beyond that, I can't say anything stronger – but compared with risks such as climate change or the next global pandemic, it isn't something that keeps me up at night.

Jacob Aron is a senior editor at New Scientist, writing about maths, physics and technology

---

Jacob's week

What I'm reading Beyond the Aquila Rift, a collection of short stories from British sci-fi master Alastair Reynolds. Great for dipping in and out of

What I'm watching

ER. Having never seen it before, it has become a daily binge-watch and a pleasure

What I'm working on Thinking about what on earth is going to happen next with AI

Waste land

J Henry Fair

Bruce Museum

DO YOU have any idea what you're looking at in this image? It might look completely abstract at first glance, but this is, in fact, an aerial photo. The white cobwebby-looking bits are bubbles, the green is algae and the straight lines are the edges of tanks.

This is the Kasslerfeld sewage-treatment plant in Duisburg, Germany, next to the Rhine river, which treats waste water from nearby businesses and homes.

“This is part of my series which illustrates environmental damage in a creative way to engage viewers,” says photographer and environmental activist J Henry Fair.

While waste-water plants help minimise the human impact on rivers, Fair points out that they don't remove all pollutants. For instance, they don't filter out all microplastics. We all have microplastics in our bodies, Fair says, but the ramifications of that aren't understood.

He has been using small aircraft to document the impact of human activity on the landscape, with the aerial view often revealing much that can't be seen from the ground. This photograph is from a new exhibition at the Bruce Museum in Greenwich, Connecticut, A Great Breach in Nature: Aerial photography by J Henry Fair.

“I first photographed it in passing, then the curators became excited about it, so I went back to reshoot it. But then could not find it. So we flew all around Duisburg, and finally saw it as we were giving up the search,” said Fair in commentary accompanying the photo on his website.

The exhibition runs until 7 February, 2027.

Michael Le Page

A complicated legacy

This biography of Stephen Hawking should be applauded for digging into various controversies – but it could have gone further, says Jacob Aron

Book

Hawking

Graham Farmelo

John Murray (UK);

Bantam (US, out 24 November)

IS THERE anything new left to say about Stephen Hawking? The world-famous physicist, who died in 2018, has been the subject of multiple books, documentaries and films, as well as penning dozens of his own tomes. In Hawking by Graham Farmelo, we now have the first authorised biography, and with it access to previously unpublished material from the Hawking family and others. Despite this, his family life is probably the least interesting part of the book – the really compelling stuff is in reassessing his later life.

Part of this is down to a problem common to almost all biographies: people and their families don't tend to do anything interesting before they become famous for the thing you actually want to read about. The first five of 17 chapters cover Hawking's life in extensive detail, thanks to detailed documents written by his parents Frank and Isabel, including an unpublished autobiography by the former. But what do we actually learn?

Unsurprisingly for someone born in 1942, Hawking had a fairly typical middle-class post-war childhood. His father enjoyed home improvement projects and sharing a telescope with Hawking, while his mother was in charge of the bulk of the childcare. The family had a near-miss with a Nazi V1 flying bomb. They grew their own vegetables to supplement meagre rations. All of this is documented and footnoted by Farmelo's impressive scholarship, and a figure like Hawking is certainly worthy of such historical investigation, but a thrilling read it is not.

Material on Hawking's diagnosis with motor neuron disease and his family's reaction to the news that he would likely be dead within a decade unsurprisingly pulls much harder on the heartstrings, as does his growing romance and eventual marriage to his first wife, Jane.

"As the decades wore on, Hawking's scientific ideas were increasingly pedestrian or indeed flat-out wrong"

It is only around a quarter of the way into the book, when Hawking

But yet again, I couldn't help feel that all of this is well-trodden ground, particularly from Jane's autobiography and the film it spawned, The Theory of Everything, in which Hawking is played by Eddie Redmayne.

ELEANORBENTALL/CORBIS VIA GETTY IMAGES

becomes a jobbing scientist, that my interest began to pick up. The physics landscape of the late 1960s and early 70s was very different to what it is today, and Farmelo does a fantastic job of immersing you in this world.

Black holes, which while still puzzling to us today are mostly uncontroversial, were thought of by many as “the cosmic counterpart of the Loch Ness Monster”, Farmelo writes – meaning Hawking was really sticking his neck out in working on them.

It is here that a distasteful recurring thread first emerges – Hawking's frequent oversight in acknowledging collaborators. At multiple times during his work in the 1970s, Hawking published papers without naming researchers who, under typical scientific norms, should have been co-authors. Farmelo highlights this, but doesn't seek to examine the psychology behind it, despite the fact there

The late Stephen Hawking taught and researched at the University of Cambridge

is a clear throughline with the fame and adoration Hawking sought and enjoyed in later life.

Still, the fact that Farmelo resurfaces these forgotten disputes, once papered over in Hawking's route to fame, as part of an authorised biography should be applauded.

Farmelo's isn't the first revisionist take on Hawking – Charles Seife got there first in 2021 with Hawking Hawking, a book that examined the making of his celebrity and failed to attract a UK publisher, perhaps because it was felt too controversial in his home country. As Hawking's life passes further into history, I imagine we can expect more.

Certainly, Hawking had some truly ground-breaking ideas about black holes in the 1970s, most notably Hawking radiation, the idea that black holes can actually emit particles and eventually even evaporate. We are still yet to detect Hawking radiation, despite most physicists believing that it exists, which is why he never won a Nobel prize – something that Farmelo explains was clearly a sore spot for Hawking.

But after that, when he catapulted to fame with the famously unreadable A Brief History of Time? As the decades wore on, Hawking's scientific ideas were increasingly pedestrian or indeed flat-out wrong.

As Farmelo points out, he frequently lost scientific bets, most strikingly that the Higgs boson didn't exist – a view Hawking advanced when he appeared as himself on The Big Bang Theory in an episode that aired just months before the particle was actually discovered.

Like many of his contemporaries in the 1990s and early 2000s, Hawking was enamoured with string theory and its variant M-theory, a framework for understanding the universe that increasingly appears today to be a dead end.

My own experience of reporting on Hawking's research near the end of his life bears this out – he would frequently come out with grabby-sounding headlines, but physicists would politely tell me off the record that they didn't feel the work was particularly new or detailed enough.

Farmelo goes some way to acknowledging this in his epilogue, but doesn't dwell on why Hawking worked in this way. I think it is probably time we did. Was it down to a lack of original ideas, his penchant for fame-seeking or his difficulties in communicating as his condition worsened, making it harder for him to publish traditional scientific papers? My honest answer, though it feels impolite to say so, is probably all three.

Where does that leave Hawking, rather than Hawking? If you have already followed his life in detail, I'm not sure you have much to gain from this book. If you only know Hawking as a celebrity scientist, Farmelo's book gets you closer to the man underneath, though I would argue Seife did it better. If it is 2042 and you're writing a centenary retrospective of Hawking's life, Farmelo's scholarship is a fantastic place to start – but I'm hoping that between now and then we can continue to decipher Hawking. When speaking of a man who famously held a party for time travellers (an incident strangely not mentioned in the book), I see no reason to constrain my recommendation to the present.

The colour of grief

A haunting space animation is a sci-fi challenge to our times, finds Davide Abbatescianni

Film Moonfish Daniel Zvereff Pending release

IN THE song Space Oddity, David Bowie's Major Tom drifts beyond the reach of Ground Control, poised between tech and terror. The protagonist in Daniel Zvereff's animation, Moonfish, occupies a similarly lonely orbit, but one that demolishes the fantasy of space travel, instead depicting work, radiation, quotas and the slow psychological erosion of being almost alone.

Cassius (Morgan Spector), a deep-space miner, is travelling with his companion Vesper (Lucy Liu) to an asteroid. Work is brutal: meet extraction quotas to earn passage home or stay, stranded. When Vesper dies in an accident, Cassius must complete the mission alone, with only her goldfish (below) and the unreliable contents of his mind for company.

Moonfish premiered at the Venice Biennale last month, and cost a mere €200,000 to make. Zvereff turns that constraint into an aesthetic: drawing in stark black and white, building from simple lines, using a boxy frame and sparse environments that

NEW EUROPE FILMS

make Cassius's universe feel both enormous and suffocating. What might have looked unfinished is the right visual language to express alienation, a man reduced to a few fragile marks against a void.

The scientific details of this world are left speculative; science ends up functioning as pressure. So, Cassius's body is another system to fail before the machinery does – prolonged exposure to radiation in space has consequences requiring medication to function. Asteroid mining isn't heroic exploration, but rather the economic logic of using vulnerable workers when tech costs more. This is one of the film's bleakest tropes – and one of its most plausible.

But its most compelling element is the bond Cassius develops with the fish, which, as grief and isolation set in, speaks in Vesper's voice. The exchanges become the film's philosophical core, touching on consciousness, memory, mortality and what distinguishes life from non-life. This could have tipped into whimsy, but Liu endows the fish with an eerie tenderness, while Spector makes Cassius's exhaustion feel physical.

Moonfish is ultimately less interested in whether humans mine asteroids than in who might, and under what terms. Its space setting amplifies existing concerns on Earth: precarious labour, dependence on tech, loneliness and the monetisation of endurance. Like Major Tom, Cassius drifts further from the systems supposed to keep him connected. The difference is that this Ground Control isn't merely distant; it is his employer.

New Scientist recommends

Jacklin Kwan

Features editor

London

When We Cease to Understand the World is a novella that explores what happens when science unravels its own core conceit: that we can understand reality.

Benjamín Labatut's characters are all real scientists, people who

push discovery so far that it can only become dangerous to them. And the picture he paints is ghastly: science will bring people to the point of real enlightenment, but that enlightenment is profoundly anti-human.

He takes impressively few liberties with discussions of science. I know, for instance, that mathematical visionary Alexander Grothendieck – a key character here – revolutionised algebraic geometry before becoming a recluse. I don't know whether his hermetic life was due to him glimpsing the "heart of the heart", the object at the deepest centre of maths. Labatut imagines this revelation as like confronting a beautiful, horrifying, many-eyed angel.

This book will haunt you.

The long and short of it

Why do modern humans have hair, anywhere? An enthusiastic and curious book looks for answers – and persuades an initially sceptical Elle Hunt

Book

Anxious and Hairy

Judy Wearing

Icon Books

PICK up Judy Wearing's Anxious and Hairy, and you might well wonder at its length: how did she possibly fill 300 pages with... hair? It is a credit to her diligent research, infectious curiosity and personable voice that, on finishing this exhaustive exploration into what the subtitle calls the “evolution and culture of human hair”, I was struck by how much I'd taken for granted.

MARTIN PARR/MAGNUM PHOTOS

From the start, Wearing sets out to win over readers who might be sceptical, taking on the broad (and, arguably, often sexist) sidelining of hair to reveal a “biological wonder” touching “on virtually every aspect of humanity”. Her own fascination dates back to her postgraduate studies at the University of Oxford, and Anxious and Hairy builds on her research into mating behaviour, especially “adaptive benefits of human hair patches”.

Why did we evolve to have tufts in some places, but not others? Even now, research remains limited. As Wearing points out: "Oddly, humans have paid more attention to the traits involved in the mating of mites and birds than the plumage of our own species."

The earliest fossil evidence of hair, she writes, dates back about 245 million years to cynodonts: warm-blooded, reptilian and the ancestors of mammals. They developed fur for insulation, but some hair follicles continued to evolve to produce milk and body odour. Homo sapiens shed their fur coats over time, but not as decisively as it seemed. The so-called naked ape, in fact, has the same number of follicles as our primate relatives, everywhere except on the soles of our feet and palms of our hands; but they produce mostly thin, fine “vellus” hairs.

The thicker, more mature hairs evolved, through sexual selection, to grow in specific locations for highly varied purposes. The hair on our head helps to regulate temperature while also offering

"At Oxford, Wearing ‘ambushed’ her peers with requests for samples from their ‘interesting bits’"

JUDY WEARING

sun protection, explaining the diversity in density, colour and curl. Curly hair, for example, blocks more solar radiation while keeping the head cooler. In men, this balance also involves facial hair, with Wearing exploring the links between bushy beards – which would have dissuaded sexual rivals – and balding.

The science behind our other furry patches is more speculative. Underarm hair might have evolved to broadcast our particular chemical blends, aiding social communication and helping to attract genetically suitable mates, she writes. As for pubic hair, Wearing's preferred hypothesis is that it evolved as a “cheating detector”, trapping fluids and odour, informing mate choice and communicating about sexual activity, though she stresses this is untested.

Even with all the unknowns, Wearing's evolutionary lens makes hair unexpectedly interesting. Culture adds an equally dense layer – and accounts for the titular “anxious”. Having too much hair in the wrong places, or not enough in the right ones,

From how you style your locks to their texture, hair is a complex business

is associated with depression, anxiety and lower quality of life.

Hair styling has been a form of self-expression since the earliest societies, and across the classes. Rameses II, one of ancient Egypt's most renowned pharaohs, used henna to hide greys. And the remains of a woman buried in the 14th century BC showed evidence of 70 coloured extensions.

Even deciding to wear hair up or down is socially mediated, and often political. Women's flowing tresses convey feminine sexuality but may be covered, bound or shorn as a means of control. Hair has also literally been a matter of life and death, writes Wearing. Natural variations, a function of genetics and geography, were used to justify racialised prejudice, abuse and enslavement. Hair texture, an official classification criteria in apartheid South Africa, determined rights and freedoms.

Anxious and Hairy reflects Wearing's careful, wide-ranging research, but the narrative is powered by her enthusiasm for hair and on-the-ground reporting. At Oxford, she made a name for herself, “ambushing” her peers with requests for samples from their “interesting bits”. Later, her quest takes her from a high-end Paris salon to Detroit for a beard and mustache convention.

She discovers that what is signified by our “patches of hair” continues to evolve. Botox, which limits eyebrow movement, may affect our ability to communicate and empathise. Overall, Wearing more than delivers on her promise to transform how we see hair.

Elle Hunt is a writer based in Norwich, UK

A wild ride through physics

The September read for the New Scientist Book Club was a challenging trip to the edges of space-time – but one that most members enjoyed, says Alison Flood

Alison Flood is head of books and culture at New Scientist, and runs the monthly New Scientist Book Club

Book

The Edge of Space-time

Chanda Prescod-Weinstein

Chanda Prescod-Weinstein

THE New Scientist Book Club has been testing its collective brain power in September with Chanda Prescod-Weinstein's The Edge of Space-Time. Subtitled “particles, poetry and the cosmic dream boogie”, this sees the New Scientist columnist and theoretical physicist set out to explain our universe, from the quantum to the inner workings of black holes.

That's a lot to get through, but as Chanda writes in her book, it is worth the effort: "We should bother with physics because it is, in part, how we as a species learn to use our minds." She went further in an exclusive essay for book club members, challenging us to get to grips with the quantum nature of space-time. "Spending time with what feels strange about physics can help prepare us to better reckon with difficult political questions," she wrote. "Given the crises we face – genocides, climate change, attacks on transgender people – we need our critical faculties now more than ever."

Her book, as member Dee55 put it, is a “wild ride”, in terms of both the physics and the in-depth conversation that members have been having about the book over on our Discord channel.

I really enjoyed it. Thanks to Chanda's friendly, encouraging and clear writing, I emerged at the end a little more knowledgeable about physics. I liked how she slipped in references to literature and popular culture to help make her points, as well as how she told us firmly: “This is supposed to be fun, and there is no test at the end.” She reiterated this in our chat: “[Part] of welcoming people into the material is saying, this isn’t something that only... a lone genius can think about. You’re

"We should bother with physics because it is, in part, how we as a species learn to use our minds"

entitled to this as well. This is your universe, too."

Many book club members were also enamoured. Jodie has been “enjoying it a lot, even if it does make my head ache at times” – I think that’s fair, Jodie! – while Dee55 found Chanda’s explanations mind-expanding: “I have [made] periodic attempts to understand Quantum Stuff and always come away fairly befuddled but also intrigued – and this time feels a little different, as if I am not exactly grasping something but perhaps catching fleeting glimpses of it.”

"I've really enjoyed The Edge of Space-time," wrote DavidCl. "It's like sitting down and spending time with a good friend who is way cleverer and more intelligent than I am while she tries to explain what it is she does, why she does it and what drives her onwards."

This book is packed with literary allusions, and with Chanda's strongly expressed political perspectives – not everyone appreciated this. Katita2026 found that Chanda's politics and her subjective takes intruded on the physics a little too much: "Dr. Prescod-Weinstein is an incredible storyteller... [but] as much as I'm enjoying the read, I find myself struggling when she tries to make the physics fit her politics."

This was a question picked over with great sensitivity by book club members. “I do agree that some of the connections are bit forced, but I think the author is trying to show not only how physics says that the universe is, but also how she from a particular background visualises that,” wrote Alan\_P.

All in all, I think this was a (challenging) hit with book club members – on to the next one.

You can find Chanda Prescod-Weinstein's essay and interview at newscientist.com/bookclubextras

---

What we are reading in October

Exit Party

Emily St. John Mandel

In October, the New Scientist Book Club is embarking on the latest novel from one of speculative fiction's biggest names: Emily St. John Mandel.

The author of flu-

pandemic novel Station

Eleven tackles quantum entanglement and parallel worlds in her latest book, Exit Party, in which our protagonists Ari and Gloria attend a party where some mysterious doppelgangers seem to be appearing.

Book club members can enjoy an extract from the novel's start, which will be available at newscientist. com/bookclubextras from

2 October, and a review from our sci-fi columnist Emily H. Wilson – she calls the book "as elegant and brilliantly executed as fans have grown to expect". There will also be an interview with St. John

Mandel later this month.

You can join the discussion with our Discord community online at bit.ly/4ewuwyo – and sign up to read along with us, if you haven't yet.

Happy reading!

Join New Scientist's free book club

Sign up at newscientist.com/bookclub

If I pass away

Canine Cared For

If I move to a care home

Canine Cared For

If I receive a life-changing diagnosis

Canine Cared For

Make sure your dog is cared

for with a Canine Care Card

You've always done what you can to protect your four legged friend, and that shouldn't change when life does. With a Canine

Canine Care Card

In the event I can no longer care for my dog(s), Dogs Trust will.

DogsTrust

Please call 020 7837 0006 Registered Charity Numbers 1167663 & SC053144. Co., Ltd 0031

Care Card, you won't have to worry about what's next for them, because we will look after and rehome them at one of our 22 rehoming centres, located all across the UK.

Apply online to get your FREE Canine Care Card at: dogstrust.org.uk/ccc E: ccc@dogstrust.org.uk or T: 020 7837 0006

Please quote "1059054", or complete this coupon:

Please send me a Dogs Trust Canine Care Card application form (Please fill out in block capitals)

Title

First Name

Surname

Address

Postcode

1059054

From time to time we would like to send you exciting updates about our work, products, services and how you can support us, including fundraising activities and research by post. If you'd rather not receive these mailings, please tick here □

We treat your personal information with care and respect your privacy. We do share your information with Dogs Trust and its subsidiaries. Your personal information is never shared with third parties for their own marketing purposes. It may be processed on our behalf by carefully selected partners. For more information about how we look after and analyse your data, please visit our privacy policy: dogstrust.org.uk/privacy.

Please send this coupon to: FREEPOST DOGSTRUSTL (No stamp required.) N.B. Service only available for residents of the UK, Ireland, Channel Islands & Isle of Man.

© Dogs Trust 2026

Registered with FUNDRAISING REGULATOR

A dog is for life®

dogstrust.org.uk

Registered Charity Numbers 1167663 & SC053144

Co. Ltd 09365971

---

Editor's pick

---

Human limitations should not constrain mathematics

19 September, p 7

From Alistair Fraser,

Whanganui, New Zealand

Terence Tao makes a persuasive case that mathematics is about more than just producing answers: discoveries must be understood, explained and incorporated into a wider body of knowledge. But I wonder whether his concern that AI may produce mathematics faster than mathematicians can absorb it contains an assumption worth questioning.

Why should the future rate of mathematical discovery be limited by the rate at which humans have traditionally been able to process it? If AI becomes capable not only of discovering mathematics, but also of verifying, organising, connecting and explaining it, then human comprehension needn't disappear – but it may cease to be the bottleneck.

Consequently, rather than diminishing, the role of mathematicians could instead shift towards deciding which questions matter, interrogating machine-generated knowledge, recognising significance and determining where discoveries should lead. The greater danger may not be that AI produces mathematics too quickly, but that we try to accommodate a fundamentally new capability within structures developed for an age when human intellectual labour was scarce.

---

New muscle-boosting drugs may be misused

5 September, p 20

From Jon Arch, Welwyn Garden City, Hertfordshire, UK

Michael Le Page's article about drugs that prevent muscle loss reminds me of the medication clenbuterol. It is used to treat asthma, but also causes a rapid, remarkable increase in skeletal muscle mass in rats and mice.

Unsurprisingly, some years after this was first discovered, several weightlifters suddenly found that they needed clenbuterol to treat their asthma. It is now banned for International Olympic Committee-tested athletes. Will the same fate await myostatin and activin antagonists?

---

Our carbon-dating efforts are safe from solar storms

Letters, 19 September

From Eric Kvaalen,

Les Essarts-le-Roi, France

Bryn Glover is concerned that carbon-14 dating may be inaccurate because the amount of this isotope in the air varies, particularly after intense solar storms. He needn't worry: it has long been known that the amount varies, and old wood is exactly the way we get out of this problem.

Over the past several decades, researchers have created a calibration curve for carbon-14 dating by looking at the pattern of thick and thin tree rings in old wood and thereby identifying when the tree lived. Then, by measuring the carbon-14, the calibration is achieved and can be used for other organic material.

---

Psilocybin could be a boon for chemotherapy patients

12 September, p 10

From Jane Giffould,

Halstead, Essex, UK

I was delighted to read that

psilocybin, the psychedelic

compound in magic mushrooms,

may one day alleviate the

numbness and tingling

experienced in peripheral

neuropathy, which I have had

for the past nine years as a result

of chemotherapy. This side effect has been growing for me, while the drugs that are meant to help clear it have had zero effect. If psilocybin works, then bring it on. I just hope it doesn't interact with the countless other chemo-related drugs to generate even more side effects.

---

Immune system's role in labour may be even greater

19 September, p 12

From Ahmad Aljesirawi,

Amman, Jordan

The finding that immune cells may kick-start contractions in childbirth is fascinating. I wonder whether there could be another connection worth investigating.

Research on human attraction has suggested that body odour may sometimes influence mate preference, including possible preferences related to differences in certain immune system genes. In simple terms, we may sometimes be attracted to partners whose immune-genetic profile differs from our own.

Could these immune differences between the mother and father – and therefore between the mother and fetus – have any relationship to the immune signalling described in this new research? For example, could recognition of fetal or paternal antigens contribute, directly or indirectly, to the activation or recruitment of maternal immune cells in the uterus near the end of pregnancy? Could this influence cytokine or chemokine signalling, macrophage activity or the mitochondrial transfer to uterine muscle cells described in the study?

I am not suggesting that this is the established mechanism, but I think it raises an interesting question: could the immune

system differences that may play a role in mate attraction also have a later biological role in pregnancy and the initiation of labour?

---

Oral flea treatments for pets may be damaging, too

Letters, 12 September

From Greg Nuttgens, Porthcawl, Bridgend, UK Discussing the effects of flea treatments on the environment, Jens Christian Jensenius suggests that oral methods may not pass through the animal. But some of the chemicals used in worming treatments for sheep and cattle, which are similar to flea treatments, do pass through the gut. They can have a serious effect on dung beetles and other insects, which feed on and break down dung, reducing biodiversity and slowing down the incorporation of dung nutrients into the soil. These chemicals should be banned for both farm and domestic use.

---

Dark matter is still elusive despite potential sighting 12 September, p 4

From Ulrich Seydel,

Melbourne, Australia

In the search for dark matter, aren't researchers putting the cart before the horse? We know dark matter has gravitational effects, but the rest is, to my knowledge at least, conjecture. Building hypotheses about dark matter and designing experiments based on those hasn't been very effective over the past few decades. Wouldn't it be more productive to stop guessing and continue observing with a wide-open mind?

---

Want to get in touch?

---

For the record

Send letters to letters@newscientist.com; see terms at newscientist.com/letters Letters sent to New Scientist, 9 Derry Street, London, W8 5HY will be delayed

Albert Einstein's general theory of relativity explains gravitational time dilation (12 September, p 22). 2.5 million animals were used in scientific research in the UK in 2025 (19 September, p 32).

THE odds of the bet Jonathan Oppenheim made in 2023 are generous, even at 5000 to 1. But if he wins, it could transform our view of the universe. The physicist is wagering that space-time, and therefore gravity, isn't governed by quantum theory. That swims firmly against mainstream physics, which says gravity must ultimately be quantum, like the rest of the fundamental forces – even if we can't currently work out how.

The wager is playful, but the question is serious, representing perhaps the most intractable puzzle in modern physics. And although Oppenheim is still out on a limb with his “post-quantum” theory of gravity, he is increasingly convinced his peers should take notice – not least because of the complexity of the alternatives. “There are all these people who think that we live in 11 dimensions and there are superstrings and space-time is emergent, looking at this theory and calling it radical,” he says. “But if you step away and think about it objectively, it’s sort of the opposite – it’s the most conservative theory you can have.”

That's not all. Accept the model's perhaps most controversial innovation, and Oppenheim's calculations show that it might elucidate another great mystery of modern physics: the nature of the invisible dark matter that seems to provide most of the gravitational oomph in the cosmos. Might he just be on to something?

For decades now, scientists have been trying to reconcile the two pillars of modern physics: quantum theory, which governs the behaviour of particles and forces on small scales, and Albert Einstein's general theory of relativity, which describes gravity and determines how the universe works on the largest scales. Each

Glitches in gravity

of these pillars has been extensively stress-tested and stood strong. Yet, according to our current understanding, the two are fundamentally incompatible.

Quantum theory explains the workings of three fundamental forces – electromagnetism and the strong and weak nuclear forces – through the exchange of quantum particles that bind matter together in different ways. But according to general relativity, massive objects generate gravity not by exchanging quantum particles, but by warping space and time around them, causing them to “fall in” towards each other.

Each theory works spectacularly well in its domain: quantum forces explain how the matter that makes up the chair we are sitting on holds together; gravity explains why our bodies fall to Earth with a bump if we miss the chair when we try to sit. But this coexistence breaks down where large scales meet small ones: when quantum particles cross over the event horizon of a massive black hole, for instance, or in the very first instants of the big bang, when our known universe began as a tiny pinprick of mass and energy.

Bringing gravity into the quantum fold has long proved tricky, but a radical approach could be the answer, discovers Leah Crane

The vast majority of physicists think that the most likely solution will be some form of quantum gravity, in which gravity is built out of tiny quantum pieces like the other forces. At extraordinarily small scales, the smoothly warping space-time of Einstein's theory would break down and pixelate, like an image zoomed in too far.

There are many different ways to quantise

"POST-QUANTUM GRAVITY CAN REPRODUCE SOME OF THE GRAVITATIONAL EFFECTS CURRENTLY ATTRIBUTED TO DARK MATTER"

gravity and space-time. Perhaps the best known is string theory, which proposes the existence of tiny, curled-up extra dimensions. Another favoured contender, loop quantum gravity, depicts a smooth space-time as emerging from a deeper fundamental network of geometric “grains”.

Oppenheim, a physicist at University College London, isn't sold on any of this. "In quantum gravity, the notion that gravity is describable in terms of space-time breaks down, but still everyone believes it's quantum rather than a theory of space-time," he says. He thinks that, if any successor theory is to reproduce the successes of general relativity, it's crucial it should maintain that theory's picture of gravity as a force acting between masses in a continuous, indivisible space-time. Post-quantum gravity does just that. "If you believe that, fundamentally, gravity is a theory of space-time, this is the theory for you," says Oppenheim.

---

Paying the price

The model and its underlying mathematics are complex, but its goal is simple: start with what we know for sure about the various quantum fields, don't quantise space-time and then see what happens. After years of painstaking work, Oppenheim has found that it actually seems to describe reality surprisingly well, so far. Crucially, gravity matches with quantum mechanics on small scales and behaves as general relativity predicts at larger ones.

In his latest paper, published in July, Oppenheim and Zachary Weller-Davies at the Perimeter Institute in Canada found that certain space-time symmetries that are crucial features of general relativity can also be applied to post-quantum gravity.

This comes at a price, however. If Oppenheim's model is to fit with quantum mechanics, it must adopt certain features of it. In particular, it must allow all objects to be in a superposition of two states at once, resolving to a single state only when the quantum wave functions that describe them “collapse” on measurement.

These states include the position of the object – but herein lies the difficulty. If an object in two positions at once produces a gravitational pull from both, this would indicate that gravity is fundamentally quantum. Yet if only one position creates a pull, this would indicate that gravity “knows” which position the object will end up at once its wave function collapses. That would violate the uncertainty principle, a key and broadly

SCIENCE PHOTO LIBRARY/ALAMY

General relativity says that gravity springs from the way mass deforms space-time

accepted precept of quantum mechanics.

The only way to resolve this, according to Oppenheim, is to introduce a degree of randomness in the form of fluctuations in the coordinates of space-time. This effect is itself small, but its consequences are significant: it would mean that, at least on very small scales, both space and time are unpredictable. The ticks of a clock, for example, would no longer be at exactly equal intervals, but wobble slightly. It is a glitch, if you will, in space-time.

Messing with the nature of time in particular isn't something to do lightly. "When we describe a system, we describe how it evolves in time, so the nature of time is a big question," says Oppenheim. Time becoming random, even in a small way, makes it very difficult to work out how we should describe systems evolving – potentially throwing a wrench into the workings of physics itself.

---

Phantom dark matter

The fact that this randomness doesn't have a known physical source troubles some researchers. “I think he's sort of going around the main thing, which is an explanation of why it is [random],” says Ivette Fuentes at the University of Southampton in the UK.

Oppenheim actually agrees. “I don’t like the fact that there are these unpredictable fluctuations,” he says. “It conflicts with the idea that everything should have a cause and we should be able to physically describe the world.” But something has to give if we are to reconcile quantum mechanics and gravity, he says. “We’ve tried all the easier approaches, and nothing has worked out, so there is a price we have to pay.”

There might be reasons not to discount the existence of the fluctuations. One comes in the shape of a foundational postulate of quantum mechanics known as the Born rule. This is a formula that gives the probability of obtaining a particular value for a quantum system's state when measuring it, and it is central to making the mathematics of wave function collapse work. The rule has held up to a century of scrutiny, but we still don't know where it comes from. In Oppenheim's model, it pops out as soon as the fluctuations in time are introduced, even though it wasn't included as a premise of the model.

What's more, in work published this May, Oppenheim and his colleagues found that post-quantum gravity can reproduce some of the gravitational effects currently attributed to dark matter. Observations suggest that dark matter outweighs normal visible matter in the cosmos by 5 parts to 1 – but we have no clue what it consists of.

“Phantom” dark matter caused by space-time fluctuations might sidestep this massive problem. This part of the model is still in its infancy, says Oppenheim: “We know that there’s an effect that looks like dark matter, but we don’t know if it’s big enough to explain anything or if it has the right properties.” He is now working to nail down the possible connection in more detail.

His peers remain largely unmoved. Carlo Rovelli at Aix-Marseille University in France is one of the originators and leading proponents of loop quantum gravity. Along with string theorist Geoff Penington at the University of California, Berkeley, he struck the bet with Oppenheim back in 2023. The exact terms are that, if gravity is proved to be quantised, Oppenheim has to buy Rovelli and Penington something worth 20p; if it is proved to be not quantised, they are on the hook for an item worth £1000.

Rovelli says the model violates both quantum mechanics and general relativity. "Why should progress be walking back on what we have understood?" he says. "It is like proposing theories in which the Earth does not rotate around the sun."

Not so, says Oppenheim. The model fits perfectly well within general relativity and, by proposing a physical origin for the Born rule, provides a necessary update to quantum mechanics. “It’s not that it breaks quantum mechanics; it actually solves one of the vexing issues that has plagued quantum mechanics since its inception,” he says.

A third view is that you simply can't make an omelette without breaking a few eggs. "When you quantise gravity, and you quantise space-time, you're also breaking general relativity," says Fuentes. "I think you have to break both relativity and quantum mechanics to unify them."

Reality will ultimately be the decider. There aren't that many models of non-quantised gravity, and the selection is even sparser if you get rid of those that aren't testable. That makes Oppenheim's theory noteworthy, says Daniel Carney at Lawrence Berkeley National Laboratory in California: he likes that there are ways to check its predictions. "This model, in my mind, is mostly useful as a test case – it says, what if we're wrong? Yeah, gravity is probably quantised like electromagnetism, but what else could happen is a shockingly difficult question to answer," says Carney. "For that, it's incredibly valuable."

Already, observations are placing some constraints on the model. For example, if it is correct, there should be random gravitational waves with a particular range of frequencies and strengths that permeate the entire universe, emanating from random space-time fluctuations early in cosmic history. Gravitational wave observatories, such as the US-based Laser Interferometer Gravitational-Wave Observatory, measure ripples in spacetime caused by the movements of highly massive objects such as black holes, and have also begun to make measurements of this weaker “stochastic gravitational wave background”.

THE VIRGO COLLABORATION/CCO 1.0

Results so far have greatly constricted this background signal's possible properties, likewise constricting the possibilities for post-quantum gravity. Meanwhile, laboratory experiments using ultracold atoms to measure gravity are providing constraints to the model at smaller scales. "There's not a lot of wiggle room, not a lot of ways to hide if there are problems with it," says Oppenheim.

---

Testing times

Ultimately, what's needed are more direct experimental tests – and recent improvements in theory and quantum sensing technologies mean they could be on the cards. The first such test would be to place a tiny mass, such as a nanoscale crystal, into a quantum superposition of two locations at once. Researchers would then try to entangle it with another crystal so that measuring the position of one instantaneously affects the result of measuring the other. Entanglement is a purely quantum effect, and so if gravity isn't quantum, it would disrupt any entanglement between the two masses. Measuring any entanglement would immediately rule out post-quantum gravity, along with all other classical models of gravity.

A second test is to subject the superpositioned mass to a gravitational field and measure its response many times in a row. If gravity is quantum, the measurement should be about the same each time. If it is

The Virgo

gravitational wave

detector measures

ripples in space-time

Jonathan Oppenheim (left) and Sougato Bose examine an experiment (below) that could help test post-quantum gravity

post-quantum, the tiny fluctuations in space-time should create inconsistencies between different runs of the experiment.

The same set-up can also be used to measure how long such a superposition lasts. “If classical gravity is coming from a quantum object, that really spoils the superposition of the mass in two places at once,” says experimentalist Sougato Bose, also at University College London – the mass will quickly collapse to a single location. Carney compares it to Schrödinger’s cat, the classic thought experiment in which a cat is trapped in a box, in a superposition of being both alive and dead until the box is opened and the cat is observed. “Instead of the cat being alive and dead, the cat will just become alive or become dead on a short time scale,” he says – with, in this instance, the cat being Oppenheim’s model.

University College London has just announced a new organisation, the David Potter Institute for Quantum Information and Spacetime, headed by Bose and Oppenheim and dedicated to testing theories such as post-quantum gravity. Actually doing the experiments is still extraordinarily fiddly, and it could be a decade or more until researchers have answers. “If we thought that gravity was quantum beyond question, then it wouldn’t make sense to even do these experiments,” says Bose. “But having these alternative

"YOU HAVE TO

BREAK BOTH

RELATIVITY

AND QUANTUM

MECHANICS TO

UNIFY THEM"

theories is a very strong motivation to test this."

“Worth testing” is, of course, not the same as “likely to be true”. Until the tests are actually performed, Oppenheim’s post-quantum gravity will probably remain on the fringes. “Does anybody really believe that the world could work the way he’s proposing? I’m sure some people do, but they would be hard to find,” says Carney.

Oppenheim sticks with the defence that his model requires far fewer assumptions, fewer tweaks to the existing models of the cosmos and the quantum world, and fewer carefully selected values and variables than many, or even most, quantum gravity models that have yet to gain experimental support, such as string theory.

Fuentes thinks that is a valid point. "We've had the last decades of just theory and theory and theory, mathematics without any real connection with what can be tested," she says. "Yes, [Oppenheim] has a crazy idea, but he proposes an experiment. This is the sort of science that we should be doing at the moment."

Leah Crane is a senior reporter at New Scientist, covering topics ranging from private space flight and cosmology to quantum mechanics

Some like it hot

We have long expected animals to move to cooler habitats as climate change bites, but it now appears that many are actually doing the opposite, finds Aisling Irwin

CLIMATE change was meant to mean big trouble for Bicknell's thrush. One of North America's rarest songbirds, the species was already struggling with mercury poisoning and habitat loss. As the bird's mountainous home warmed, ecologists worried that a new threat would move in: the American red squirrel, a mammal that will supplement its seed-rich diet with songbird eggs and nestlings. The squirrels are found downhill of the thrushes and were expected to move up the mountainside as temperatures rose.

But that didn't happen. Toni Lyn Morelli, an ecologist at the University of Massachusetts Amherst, discovered that the squirrels have actually begun moving downhill. It is good news for songbird conservationists, but puzzling news for ecologists more generally. The squirrels seem to be heading the wrong way.

They aren't alone. In some locations, as many as half the species studied are defying the prediction that they would shift to cooler places. The discovery is making it difficult for conservationists and land managers to protect vulnerable species. Working on the assumption that the species would move away from the heat, they have been conserving or buying up cooler areas they assumed would become climate refuges. But what if they are protecting land in the wrong places?

"I am left with more questions than answers about what is the best pathway to safeguard things in the face of climate change," says Inna Osmolovsky, an ecologist at the University of New South Wales in Australia.

Researchers are now going back to basics in order to understand what is going on, and to devise conservation strategies that are a better fit for our warming world.

For most of the past 30 years, ecologists have assumed that climate change-driven range shifts would be largely one way: into colder places. The first good evidence for this came in 1996 when Camille Parmesan, then at the University of California, Santa Barbara, demonstrated that the Edith's checkerspot butterfly in western North America was quitting the increasing warmth of its southern, low-elevation habitats for the cool of higher and more northerly climes.

GREGORI SAAVEDRA

By 2003, Parmesan, along with Gary Yohe at Wesleyan University in Connecticut, was even quantifying the rate of the shift. Using data from hundreds of animal and plant species, they calculated that, every decade, species were moving an average of about 6 metres uphill or 6 kilometres towards the poles.

That rate was later revised upwards. In 2011, with more data to work with, ecologist Chris Thomas and his colleagues at the University of York, UK, calculated that species were, on average, actually moving 11 metres further uphill or 17 kilometres poleward each decade. The scientists also found a link between the degree of warming particular species experienced and the size of their shift –

the first evidence, they said, of a causal link.

The study caused a stir, but as Thomas pointed out at the time, within the data lay a more complex story. Although species from alpine chipmunks to British butterflies and from red foxes in the Arctic to boreal spruce and pine trees were moving to cooler places, a quarter of species tracked in the analysis were heading downhill, not uphill, and more than a fifth were shifting towards the equator rather than the poles.

With additional data gathered since then, Morelli says it now seems that “less than half of the species” have moved into cooler environments. The big question is: why are so many species apparently travelling the wrong way?

---

Embracing the heat

The answer is, in a word, complicated. A particularly baffling observation is that seemingly similar animals living in the same habitat are responding to climate change in completely different ways. For instance, about 20 years ago, Steve Beissinger, an ecologist at the University of California, Berkeley, and his colleagues climbed up and down 3000 metres of elevation in Yosemite National Park, California. They wanted to discover how the ranges of 28 small mammal species have shifted since being recorded in the early 20th century – when the national park was about 3°C cooler – by a far-sighted biologist named Joseph Grinnell.

“Half the species showed movement upslope, but the other half didn’t move at all,” says Beissinger. “Two shrews had expanded their ranges downslope.”

Beissinger then moved on to studying birds in the Sierra Nevada mountains in the western US. "And then it got a lot more puzzling," he says. He and his colleagues discovered that discrete populations of the same bird species were responding differently to climate change. The white-headed woodpecker, for instance, was moving uphill in the north of the Sierra Nevada, downhill in the middle and nowhere in the south of the mountain range.

“We scratched our heads about this for a long time,” says Beissinger. They eventually concluded that temperature change alone is too crude a predictor of animal behaviour.

They now realise that changes in snow and rainfall are important, too. For instance, although shifting downhill may involve moving into an area that is uncomfortably warm, heading uphill might mean moving into one that is uncomfortably wet. If the cost of all that precipitation is deemed too high, a population might move downhill as the least bad option.

STONE NATURE PHOTOGRAPHY/ALAMY

It might seem obvious that you need to take precipitation into account when predicting how species will move, admits Morelli. Ecologists are aware that “multifarious complications” affect where a species chooses to live, says Thomas, but they were trying to keep their forecast models simple – and factoring in rain and snow makes them much more complicated. Precipitation changes are “all over the place”, says Morelli. “It doesn’t have the same clear direction [as temperature].”

This isn't the only factor that can complicate predictions, as Beissinger found in his next study. The Mojave desert, in the south-western US, is home to a riot of small mammals – voles, chipmunks, deer mice, kangaroo rats and the like – as well as 200 species of bird. It has warmed by $2^{\circ}\mathrm{C}$ over the past century and can sometimes top $45^{\circ}\mathrm{C}$ (113°F). Beissinger found that, as expected, bird populations there were “collapsing”. But the mammals were doing just fine.

MINDEN PICTURES/ALAMY

The reason, he and his colleagues discovered, was that the mammals shelter from the fierce heat of the day in their burrows, while the birds air-fry overhead. “The lesson we took away from that was that even at the same location, species could be responding differently to the same level of climate change,” he says. It is all to do with their actual exposure. The models were going to have to get a lot more complicated if they were to accurately predict how ranges are shifting.

So, biologists have begun throwing everything they can at the problem. They have built massive databases of traits, in which they have collected all kinds of details about how organisms behave, from mobility to dietary preferences. They have studied range shifts all over the globe. The holy grail, says Monte Neate-Clegg, an ecologist at the University of California, Davis, is that somewhere in the data will be clues to a formula that can successfully predict where any particular species or community will move, and so indicate where it would be best for humans to intervene.

The approach seemed promising, says Beissinger. In particular, there was an expectation that organisms that live fast and die young would be among the most successful at shifting upwards and polewards. Such species produce a lot of offspring and

White-headed woodpeckers (above) respond differently to climate change depending on where they live; Ord's kangaroo rats (left) can escape heat in burrows; Edith's checkerspot butterfly (below, left) was an early indicator of species moving to cooler habitats; American red squirrels (below) are moving downhill in search of food

RICK & NORA BOWERS/ALAMY

BLICKWINKEL/ALAMY

“Even at the same location, species respond differently to the same level of climate change”

tend to tolerate a wide range of conditions, such as temperature and wetness, he says. They are “better at pushing at the margins of their ranges”.

To test this expectation, Beissinger and his colleague Eric Riddell at the University of North Carolina at Chapel Hill analysed all the studies that have tried to explain range shifts using behavioural traits. But they concluded that no reliable predictions can be made using such data.

“They don’t really work super well,” says Lise Comte, an ecologist at the California-based non-profit organisation Conservation Science Partners. “And so, the community overall is a little disappointed.”

To put it another way, researchers are still no nearer to identifying that holy grail formula. In fact, there are so many complicating factors influencing an individual species' behaviour that there may be no simple way to predict how it will respond to climate change – even though, across all species, you can detect that upwards and polewards signal.

Osmolovsky, who studies the movement of alpine plants in Australia, thinks more attention should be paid to how climate change alters interactions between species: "We're thinking [about] just this one plant that is doing its own thing in response to climate change," she says. "But, actually, whole communities are responding and maybe the interactions are also impacted."

---

Expect the unexpected

For example, research from last year showed that microbes that are typically harmful to plants can become mutualistic in more stressful environments. Osmolovsky says that “unpredictable friendships” like these can make life easier for a species even as its climate warms.

Beissinger reels off the list that scientists now need to consider when trying to predict how a species will respond to climate change. "Competition, disease, predation and human impacts like land-use change," he says. Moreover, he adds, species may reduce or even eliminate the need to move by changing their behaviour – breeding earlier in the year, for instance.

Although recognition of this complexity makes it hard to predict how a species will respond to climate change, Neate-Clegg says there are two practical things that can still help conservationists.

The first is to connect habitats together, so that organisms can move up, down, north, south or even east or west, according to how they can best cope with climate change. “The worst thing you can do is fragment these landscapes that species are trying to shift through,” he says.

The other is to enhance the complexity in a habitat, so that species have options for adjusting their behaviour as the climate changes. This means “not just having a forest, but having a complex forest that might provide more options for buffering”, says Neate-Clegg. Lots of different species and ages of tree, for example, mean all sorts of options for shade.

Beyond this, recognising that there is no universal model that can predict each species' response to climate change also reminds ecologists that successful conservation efforts often require a detailed study of the species in its local habitat.

“Conservation has always been this multi-pronged approach of ecosystem protection by protected areas and then species-level conservation. And I think both avenues could work,” says Neate-Clegg.

Taking this approach can even help explain why the American red squirrels didn't move into the territory occupied by Bicknell's thrush. It turns out that a fabulous new habitat has emerged further downhill. For decades, logging and acid rain had wiped out the spruce trees that once grew there and that produce the squirrels' main food source, cones. Today, both of those pressures are much reduced, and spruce trees are growing at lower elevations again.

In the end, says Morelli, “they are just this little mammal trying to make do in feet and feet of snow”. Climate change may seem to offer incentives for the squirrels to head uphill, but the best answer for them is to move in the opposite direction, and follow the cones.

Aisling Irwin is an

award-winning journalist

focusing on the environment

CAMERA PRESS/ANDREW CROWLEY

"The old-school psychiatric approach of using drugs to alter brain chemicals is just a tiny piece of the conversation"

Therapy and drugs are the go-to options for improving mental health, but psychologist Kimberley Wilson tells Olivia Goldhill why what we eat also has a crucial role to play

"Nutrition changes our brain's ability to process thoughts and emotions"

D IET is key to healthy living but, strangely enough, the topic rarely comes up in conversations about mental health. Instead, we tend to focus on medication and therapy, and largely ignore the food that we eat – which, according to psychologist Kimberley Wilson, is a huge mistake.

After working at what was once Europe's largest women's prison, HM Prison Holloway in London, she became keenly aware of the powerful influence diet has on mental wellbeing. She saw how, deprived of a balanced diet and nutrients, it was so much harder for prisoners to control their emotions. And she realised that proper nutrition is just as important for managing mental health as medication and therapy.

She spoke with New Scientist about why nutrition is the foundation of brain health and how the foods we eat can treat mental illness.

Olivia Goldhill: How does nutrition influence brain health? I think most people understand that food is important for our bodies but are less familiar with its influence on our minds.

Kimberley Wilson: Mental health is an emergent property of a physical organ. In the same way that your heartbeat is an emergent property of your heart, all of these things that we think about as mental wellbeing are an emergent property of your brain. If we want a coherent psychological experience, we should be thinking about the physical health and structural integrity of the brain as our starting point – and nutrition changes our brain's ability to process thoughts and emotions, just like it influences the health of any other organ.

Why do you think we so rarely consider this mind-body connection? I think it is Cartesian. Four hundred years ago, René Descartes made the argument that there is an essence of the self that is separate from and superior to the physical body. It created such an impression on medicine and then psychiatry that we started approaching mental health as a neck-up intervention – the rest of the body is completely irrelevant.

The brain is also out of sight. Weightlifting changes your body very quickly, often within a few weeks. But because you can't physically see changes to your diet, exercise and sleep in your brain, you have to wait for a subjective shift that can take six months.

There is also a bit of egocentrism. We think of ourselves as deeply complex individuals. We don't like the idea that malaise might be improved by a change of diet and a good night's sleep. We seek either a complex pharmacological compound or a deep-probing-type therapy to get to our core complexities.

Obviously, I provide some of that, too. But we also know that profound changes can be made with nutritional and lifestyle interventions. I see that in some prisons, where randomised controlled trials showed that changes in nutrition reduced violence. That is some of the starkest, clearest evidence that the brain is a physical organ, and if you want it to function well, you need to feed it properly.

You worked in Holloway prison and young offenders institute up until 2013 – it is now closed, but was then the largest women's prison in Europe. I imagine your experience there shaped your approach to mental health.

In prison, you are working with a very traumatised and traumatising population. The focus is on the emergent behaviours – this person is angry, they are aggressive, they are dangerous, they are frightening – and most people aren't thinking about the machinery from which those behaviours are emerging.

But we are dealing with a group of people who have experienced extreme deprivation on all levels: emotional neglect, often abuse, but also severe nutritional, and often substance-related, harms to the brain.

We have a justice system predicated on volition: you know right from wrong, and if you choose to do the wrong thing, then you deserve the consequences. But that is based on the assumption that the decision-making machinery is working properly.

This is where we get into almost philosophical grounds with prison violence. If I get into a car and crash, and I didn't know the brakes were shot, am I entirely responsible for the accident? Or are there mitigating circumstances that need to be taken into consideration?

In terms of brain health and prison violence, we see that when you improve nutrition, violence drops by 30 per cent. These were supplement trials of broad-spectrum micronutrients, so vitamins, minerals and, in some cases, omega-3.

Nobody wants to let people off scot-free, but I think we need to be thoughtful about the assumptions we make about behaviour, not just in prison, but across the board.

---

Speaking more generally, what impact do physical health and diet have on our mental wellbeing?

There is lots of evidence around the physical health of either the whole body or the brain impacting mental health. We have good data showing that 1 gram per day of omega-3 – where 60 per cent is EPA [a fatty acid found in omega-3] – can reduce depression. It is thought that inflammation drives depressive symptoms for some people. Omega-3s are the necessary precursors of inflammation's “off switch”.

Then there is data looking at brain trajectories and psychopathology. Brain scans have shown us that if the brain was smaller than expected in the prefrontal and parietal lobes, which manage higher-level thinking, that was associated with increased risk of psychopathology. We are talking about the physical structure of the brain as a risk factor.

Iodine deficiency in the perinatal period and early infancy, according to the World Health Organization, is the leading cause of preventable brain damage worldwide and is associated with low IQ.

RAUL ARBOLEDA/AFP VIA GETTY IMAGES

Nutrition has a direct influence on inmates' mood, emotions and behaviours

---

NewScientist Live

Hear more from

Kimberley Wilson at this

year's New Scientist Live.

On 10 October, she will

explore how to build

lifelong brain health, delving into the key nutrients that support neural development, cognitive performance and emotional resilience.

For more information, visit newscientist.com/nslmag

Some of the most robust data on depression prevention is around exercise. But we rarely think about this because [most people] think that mental health needs to be addressed with either psychotherapy or drugs.

---

Traditionally, psychiatrists and psychologists are very brain-focused, but your approach to the brain seems to be more holistic.

Yes, and I think the old-school psychiatric approach of using drugs to alter brain chemicals is just a tiny piece of the conversation. These compounds – serotonin or dopamine – are important for regulating mood, but they are also made from nutrients.

If you are anaemic, you aren't going to have good serotonin production because the brain needs iron for this process, and we see associations between iron deficiency and mood disorders like depression. One of the simplest fixes for a mood disorder is to check, are you anaemic? If so, we can correct this with either diet or iron supplementation. Then, we have improved your fatigue, your flatness of affect, and you haven't had to go to therapy for six months.

How do you think data on nutrition and mental health should inform diagnosis and treatment? What needs to happen is a better assessment process. Rather than a binary “are you depressed, are you not”, what we want to ask is, what is driving your depression? Is it your childhood trauma, in which case, let’s fast-track you to in-depth therapy. Is it relationship stress? Is it workplace demands? Is it that you have eaten only two chocolate muffins in the last four days? That isn’t going to help you feel good. If we intervene there first, then we can make a difference in the way that you feel very quickly.

So, if you lived off only ultra-processed foods, for example, what would happen to the brain? The brain is the hungriest organ in the body. We need to supply it with the nutrients it requires so that we can feel good, focus, have goal-directed behaviour – all of that stuff.

A single doughnut isn't going to meaningfully hurt your brain. But the more your diet is made of these foods, the lower your nutrient status. You will be lower on things like [vitamin] B3 and B12, omega-3, selenium and zinc. That is a problem, because your brain is then depleted of the things it needs to function well.

We've talked about the nutritional effects of diet on the brain, but I'm also curious about the sociological importance of eating. Are there health benefits to sharing our meals with others? I think so. I think cooking and eating together is part of what makes us human. Across human evolution, eating together was a necessity, because you couldn't go out and hunt and gather for yourself. It was a group process: I share what I have, and you share what you have.

There is lovely data on how we feel more trusting of people when we eat together, and it's not simply that you eat with people you already trust. There is something about eating that builds those relationships, and I think that has to do with the physiology of eating. You have your fight and flight, and then you have your rest and digest. You need to be calm, you need to feel safe [and] you need to feel like you are with people who are trustworthy, so that you can drop your guard and digest your food.

You just had a baby. How has pregnancy and the postpartum period influenced your understanding of the importance of nutrition, if at all? I took the position that I am building a brand-new brain from scratch. So, it was important for me to think about what I am supplying that brain with in order for it to develop optimally.

I was comparing prenatal supplements with research on what nutrients are important for brain development and just finding them massively wanting. Basically, no prenatal

ALEXANDRKOLESNIKOV/GETTY IMAGES

A diet rich in vitamins and omega-3s can improve brain health

PAUL FAITH/AFP VIA GETTY IMAGES

vitamin has enough of the nutrient choline, simply because choline salts are bulky. Though we know that choline can do the same thing that folic acid does in reducing the risk of neurological congenital conditions, and we see that choline supplementation during pregnancy improves attention in children at age 7. I am breastfeeding, and those nutrients become present in your breast milk. So, I take a choline supplement, to make sure that I am getting enough for my son.

---

Is there one nutritional change for mental wellbeing that you think most people could benefit from?

I would probably say increasing your intake of omega-3 – ideally by eating at least one portion of oily fish a week, and vegans or vegetarians may want to take a supplement. That is where some of the strongest data is, showing reduced risk of depression and suicidal ideation, we think as a function of reduced neuroinflammation. Given that hardly anybody in the UK is getting enough fish, we can assume that a good number of people would be helped by just making that one change.

---

Do you think we will see change in how people talk about diet and mental health?

When I first started talking about this research, I was speaking to a colleague and he just laughed at me. It was quite disappointing, and he just said, well, I have never heard of it.

It's hard to tell if things are getting better because I don't know whether I'm now in a bit of a bubble. I hope the old approach is being pushed out by more open-minded, grounded-in-the-new-research, optimistic individuals saying, actually, the brain is a physical organ and there is something we can do to improve mental health.

Olivia Goldhill is a features editor at New Scientist focused on biomedicine, health and neuroscience

The perfect gift for curious kids!

√ A brilliant monthly magazine for ages 7–14

√ Amazing facts, photos, and screen-free fun

√ All their favourite topics, from animals, space and nature to art, sport, science and inventions

√ Puzzles, quizzes, games, activities and jokes

√ 10 major awards, including Children's Magazine of the Year

√ Help them fall in love with reading and learning

OFFER ENDS 31 OCT

SAVE 60%

INCLUDES BRITANNICA KIDS' ENCYCLOPEDIA WORTH £27\*

---

Quick crossword #218 Set by Richard Smyth

Scribble zone

Cryptic crossword

next week

---