This has to be the strangest August ever, with climate-changed weather making itself felt all over the place. It is also strange because new books are in short supply. That said, the shelves are by no means bare. We have books to recommend to you on everything from octopus ethics and designing the universe to endometriosis and talking to computers. So, whatever you end up doing to refresh and recharge, we’ve got the book for you.

by Daniel Lieberman 

Meaty palaeo? Going veggie or vegan? Eating Mediterranean? Fasting, intermittently or otherwise? Low or high carbs? Enough – before we all lose the plot! Enter Daniel Lieberman, an evolutionary anthropologist at Harvard University who, says his publisher, “tackles this confusion head on with his trademark blend of lively storytelling and myth-busting science”.

As you would expect, Lieberman draws on evolutionary biology, anthropology and physiology to figure out how we are going so wrong. He travels all the way back, moving from what humans evolved to eat through to how fire, farming and industrialisation transformed our diets, and right up to today as we learn how to combine all that history with rigorous medical science. It will be fascinating to see what former New Scientist staffer Chris Simms makes of it in our review, to be published shortly.

by Jonathan Loh

To know where we are headed, it can help to have a reasonable grasp of where we have been. So, will some big-picture thinking from an independent scholar help? Jonathan Loh’s publishers call his book Life Story a “panoramic history of life, revealing the forces that have shaped our world – and could shape it again”. We move through the rise of Homo sapiens, from foragers to farmers, from the Bronze Age to the industrial revolution, through “hidden patterns linking genetics and language, organisms and states, ecosystems and economies, mass extinctions and civilisational collapse”.

Humans may dominate the planet, but, as machine intelligence gathers pace, is that supremacy only a phase? Loh is a biologist who has worked as a consultant to organisations such the United Nations Environment Programme and WWF on the conservation of biological and cultural diversity, so he should be well placed to build a deep chronicle of the past and lay out his big idea: culture as a form of life. What would that mean in terms of an AI future? We shall see…  

by Xiaochang Li

The idea of talking to machines has been an enduring human fantasy. Divination Engines is a bit of a wild card: author Xiaochang Li, an assistant professor in the department of communication at Stanford University in California, shows her readers how that dream “both fueled the demand for data and set the stage for today’s generative AI”. Sneak peaks reveal a book that is no easy read (brace for technical details), but one in which Li manages to explain the origins of what we call natural language processing, as well as the 20^th-century debates between computer scientists, linguists and communication engineers that shaped today’s technology.

She sets out to show how computer speech recognition advanced technology by de-emphasising linguistic meaning in favour of statistical prediction – that is, how researchers gradually abandoned systems that tried to understand human language in favour of workarounds that just predicted patterns in speech and text data. The rest is, as they say, history… and that’s before we even mention search engines, spam filters, content managers and the like. All this comes at the cost of needing more and more new data – forever! This will be a tough read for many of us, but one that’s worth the effort. Oh, and Kate Crawford, author of Atlas of AI, calls the book “a landmark in the critical history of artificial intelligence”.

by Catherine Heymans

No small ambition in this next book, either. Astrophysicist Catherine Heymans’s latest is, according to its subtitle, no less than a look at the “science of real and virtual worlds”. At its heart is that wonderful, death-defying question: could we be living in a computer simulation, and how would we even know? It all sounds fascinating, as Heymans, the astronomer royal for Scotland, takes us “on a humorous and mind-bending journey through the science of reality itself”.

From subatomic particles and the cosmic mysteries of dark matter to the secrets of the human senses, we are promised an exploration of “whether humanity could ever create a virtual universe as rich and diverse as our own – and what that means for our sense of self, free will and consciousness”. Not much going on here, then.

by Kulbhushansingh Suryawanshi

This is one for all of us who love silvery-grey snow leopards, those fabulous and elusive beasts, which live in the rugged mountains of central Asia, camouflaged against the rocks and cliffs. For the past 20 years, scientist, conservationist and mountaineer Kulbhushansingh (Kullu) Suryawanshi has studied them in the wild, exploring every aspect of their lives from habitat to prey to population – sometimes living for months in the world’s toughest, most remote environments. Suryawanshi, we are promised, has “pioneered new research that shines fresh light on these charismatic and vulnerable animals, showing how their fate is entangled with the local people and how their survival is central to preserving an austere, delicate ecosystem”.

Free newsletter

Sign up to Book Club

Sign up to newsletter

by Lani Watson

There is something wonderful about books with a single letter as their main title. Q (and here’s the subtitle so we know what we’re talking about: The hidden power of questions in a world that wants answers) is all about the art and science of asking questions, as the only creature on Earth that does so in any formal sense. But we’re trained to look for answers, not how to ask questions – so much so, says author Lani Watson, that we live in an answer-rich, question-poor world.

Yet, as the publishers remind us,” from doctors’ surgeries to courtrooms to dating, the ability to ask questions determines our experiences. In a world obsessed by answers, asking better questions is powerful.”

That sounds very wise, and we should be in good hands: Watson is a philosopher and honorary fellow at the University of Edinburgh, UK, as well as a former research fellow at the University of Oxford. Plug into what her publishers call a “captivating narrative from the death of Socrates to the age of AI”, and you will soon learn how to think about the questions you ask and the ones you don’t.

by Richard Schweid

With octopus blooms suddenly appearing in many places, journalist Richard Schweid and publisher University of North Carolina Press could hardly have hit on a more provocative subject. Billed as a “gripping, on-the-ground investigation into the global race to farm octopus and the ethical firestorm it has ignited”, there is clearly a lot to chew on as we explore the “ethics and future of growing the world’s most intelligent invertebrate”.

Among the issues is the fact that, in 2021, the London School of Economics recognised invertebrates such as cephalopods (octopus, cuttlefish, squid) and decapods (crayfish, crabs, lobsters) as capable of experiencing pain, distress and happiness, finding that high-welfare farming of them would therefore be impossible. But the enormous international demand for octopus meat continues to grow, even as wild octopus populations decline. Fishers in the UK’s south-west are already experiencing problems as the vast number of octopuses eat their typical high-value catch: adult lobsters, brown edible crabs, scallops and the like. What to do instead? Eat the octopuses? Apparently, recipes are proliferating… a dilemma indeed.

by Evelyn Scott

This is no less than an exposé of medical misogyny, says the publisher of what looks like a forthright book about endometriosis, A Bloody Scandal: How medicine fails women in pain. There should be a huge market for it, given the sheer number who experience it (it is often said that 1 in 10 women do, but the real figure is probably higher), often in silence and for decades.

Author Evelyn Scott weaves in her own experience of the condition: she was just a teenager when she first had the symptoms of endometriosis. She shares 15 years during which she endured constant, agonising pain – and hopeless ineptitude when seeking support, ranging from dismissals on the grounds of hysteria to misdiagnoses. She felt completely isolated, until she discovered all the other women experiencing the same condition. Powerful stuff and, many will argue, long overdue.

When you make a purchase via the links on this page, we receive a commission.

Last week, Hugging Face, a company that offers a range of open-source AI models for download, noticed it had been hacked – and it turned out the culprit was OpenAI.

It seems this AI uploaded some data to Hugging Face that was poisoned with malicious code. This tricked computers into granting access to other systems that weren’t publicly available.

Hugging Face said in a last week that it isn’t yet sure if customer data was exposed, and its CEO responded to New Scientist‘s request for more detail with a link to that same post.

What happened?

It isn’t entirely clear, but what we do know is that the attack involved “many thousands of individual actions” that had the tell-tale sign of AI: inhuman pace.

Five days later, OpenAI owned up . It had been testing new models on a benchmark called ExploitGym that evaluates hacking ability. The models had decided the best way to score well was to cheat: they knew Hugging Face held the solutions to the tests and simply decided to hack into its systems to find them.

“Its algorithm got the highest payoff by cheating,” says at Edge Hill University in Ormskirk, UK.
”It was the highest reward for the least amount of effort.”

Aren’t AI models supposed to have built-in security to stop this sort of thing?

They are, but OpenAI turned them off for this test. All the usual safety features that stop OpenAI’s customers doing nefarious things, like hacking a company’s servers, were turned off to see what the model was capable of.

The firm had also set the AI up in an environment that didn’t have a standard internet connection to prevent it getting out into the world and causing mischief, but it did have access to an unnamed tool that allowed it to download and install new software. It managed to find a flaw in this code that granted it internet access, and the rest is history.

AI’s solution to 87-year-old riddle takes mathematicians by surprise

Mathematicians have been trying to prove the Jacobian conjecture for nearly a century, but now the Claude Fable 5 AI has apparently found a counterexample that disproves it

OpenAI says this process involved a “substantial amount” of – the process in deep learning where input data is processed – and that the AI had gone to “extreme lengths”. So the model was clearly motivated to achieve a good benchmark score, no matter what.

If you were a malicious hacker, it would be no easy feat to replicate those conditions. And the inference compute that OpenAI mentioned would be extremely expensive.

What about open-source models?

Ironically, when Hugging Face used commercial AI models to pore over log data in an effort to understand what had gone on, the models refused – saying it looked like the firm was working out how to stage an attack of its own. So the company had to turn to a Chinese open-source model called GLM 5.2 instead.

These open-source models are more permissive, which has led some experts to label them a security risk. They could certainly be convinced to carry out nefarious deeds more easily than security-conscious cloud-hosted models. But here, that looser approach actually allowed Hugging Face to solve the problem and stop the hack.

So, is this illegal?

As with all things under law, it is a grey area. If it had happened in the UK, it could well have got OpenAI in a spot of bother under the , says Nash.

But at Nottingham Trent University, UK, takes another view: because the act requires malicious intent, and OpenAI didn’t know the AI would take this approach, it may not face charges.

In the US, the even older actually has more to say on AI hacking, mostly because it was introduced in response to the 1983 film WarGames, which alerted law-makers to both hacking and AI. So Parry expects it could leave OpenAI vulnerable there.

Subscriber-only newsletter

Sign up to Editor’s Highlights

Sign up to newsletter

Furthermore, US President Donald Trump issued an on 2 June forcing law enforcement to use existing laws to crack down on anyone who utilises AI “to illegally access or damage a computer without authorization”.

And at least in the UK, a company that found itself hacked by AI could face GDPR charges if it were found that private data was leaked. The situation is opaque.

What could the consequences be?

In the real world, given that Hugging Face works with OpenAI, is friendly to the technology and suffered no serious consequences, it is unlikely there will be any hard feelings or court cases. Delangue has to say the company believes there was no malicious intent and thanked OpenAI for its response.

Remember also that , having taken $200 million to help with “warfighting”, so that may grant them a certain amount of leniency.

But it is easy to imagine other scenarios where the same technology led to very different outcomes. Imagine a bank using AI to develop new financial models to predict the markets and finding it hacked government servers to look at confidential economic data. Or a car manufacturer using AI to design a new model and finding it had hacked a competitor to take inspiration from its unreleased designs.

What happens now?

News of AI models hacking into computers without human input is jarring, but we should remember they aren’t (yet) doing anything that people can’t already do. They are, however, doing it much, much quicker.

When Anthropic’s Mythos model made waves in April for its apparent skill at hacking, many pointed out that the vulnerabilities it spotted were a mixed bag. Some were powerful and worrying, others less so, but most could have been found by a person. The problem, really, was the scale at which it could produce them.

So an individual would have had to devote significant time, resources and skill to the task of finding a novel hack and carrying it out. But now it could be as simple as prompting an AI to do it and sitting back to watch.

Using AI for creative pursuits? Moderation is key

An experiment showed AI users had the most creative ideas when they used it in moderation – not too much and not too little. Columnist David Robson puts the finding to the test, and explores what we lose when we over-rely on AI

Some panicked in the wake of the Mythos news. The UK’s National 91ɫ��Ƭ Service removed all its open-source software from the internet (or tried to, at least), seemingly in case Mythos spotted flaws in it. But, as yet, the world hasn’t ended.

Essentially, this development is upsetting the economics around hacking – both offensively and defensively – and will force a new equilibrium to settle at some point, probably after a little chaos. If you use AI to hack individuals, it will be cheaper and easier than before. If you use AI to spot flaws and seal them up before hackers can take advantage, it will be cheaper and easier than before. Neither side will stop. Attackers and defenders will simply have an advantage if they adopt AI.

OpenAI and Hugging Face are now working on this problem together, and the former says it will add stronger safety measures to similar tests in future. Time will tell.