A new book warns that the school system may be “broken beyond repair”, claiming that it is deepening inequality and making children ill. We are crying out for systemic transformation: a completely new vision of what education involves, however challenging that may be Hilary Cremin In Rewilding Education, Professor Hilary Cremin argues that modern schooling is defined by an obsession with standardisation and outdated thinking, while it fails to nurture creativity, critical thought, or the physical and mental health of students and teachers. Cremin, who is Head of the Faculty of Education at the University of Cambridge, draws on decades of experience as a teacher, academic and consultant – as well as the work of other scholars – to put forward a programme for “long-term, radical change”, including a stronger focus on students’ social and emotional development alongside academic achievement. The book’s numerous proposals include more lessons outdoors, and more projects that connect students to their communities beyond the school gates. Steps such as these, she argues, would help prepare young people to live responsibly – and well – in a rapidly changing world. Cremin acknowledges that these ideas may be disparaged by traditionalists and policy-makers – as, indeed, they have been before. In 2013, she was one of 100 academic critics of Michael Gove’s educational reforms whom the then Education Secretary branded “enemies of promise”. More than a decade later, she argues, there is still no evidence that those reforms, like many before and since, have narrowed the attainment gap between wealthy and poorer students as promised. Research shows that the gap widens throughout school, reaching the equivalent of more than 19 months of learning by the end of secondary education. “Despite decades of reform, I think the school system as we presently configure it may be beyond redemption,” Cremin said. “This isn’t an attack on the idea of education, or on the thousands of brilliant teachers who give the job their all. But government after government has tinkered with education when the basic model is obsolete.” “If we keep preparing children for the second half of the 21st century using a system designed in the 19th, it could do catastrophic harm. We need to rethink what it means to educate, and what we are educating for.” Rewilding Education challenges the ‘myth of social mobility’, arguing that education functions more as a sorting mechanism than a levelling force. High-performing school still admit disproportionately few disadvantaged young people, and poverty remains the strongest available predictor of student outcomes. The chimerical belief persists that good grades will secure students a better future. “None of the ideas driving schools policy really stands up to scrutiny,” Cremin writes, “yet this hardly seems to matter”. Cremin contends that schools often resemble outdated, factory-style production lines: rigid, standardised and with sometimes militaristic discipline. This, she suggests, suppresses curiosity, discourages critical thinking and disempowers teachers. Her critique of the effects on physical and mental health is particularly urgent. Cremin argues that schools are making students and teachers ill. She presents evidence linking the loss of physical education and the sale of school playing fields to rising childhood obesity, and notes that even basic needs – such as access to adequate toilet facilities – often go unmet. High-stakes testing, she adds, is fuelling poor mental health, while zero-tolerance behaviour policies have driven a 60% rise in permanent exclusions since 2015, with disadvantaged students four times more likely to be excluded. Students and teachers, she suggests, sometimes turn to medication to cope with an “ailing system”. This bleak reality, she argues, demands more than incremental reform. The book calls for a new educational model for a new kind of future – one shaped by the climate crisis, downward mobility, Generative AI and post-truth politics. “We are educating for jobs and lifestyles that will soon cease to exist,” Cremin writes, “while failing to educate for those that don’t yet exist.” This leads Cremin to call for education to be ‘rewilded’ – a metaphor drawn from ecological restoration. In schools, it implies letting go of rigid, one-size-fits-all structures, and allowing less predictable and more holistic forms of learning to emerge. Nature plays a central role in her vision. Drawing on thinkers like Rabindranath Tagore, Cremin argues that schools should treat the natural world as a “co-educator”. She encourages outdoor and experience-based learning and suggests that even small changes – like planting trees, creating school gardens or nature-inspired arts activities – could help foster greater respect for the environment. Rewilding Education also urges a rebalancing towards project-based learning, the arts and civic engagement. Students, Cremin argues, must learn not only to reproduce knowledge, but to act with wisdom and care, and to think critically about complex problems. This requires education for “body, mind, heart and soul”. She proposes, for example, giving students time to walk and reflect when grappling with difficult questions, and highlights research linking later start times for adolescents – who have different sleep patterns – to better performance and wellbeing. She also champions mindfulness and ‘metacognitive’ approaches, that help children reflect on how they are thinking while they are learning. In a chapter Cremin anticipates critics will deliberately misread, she calls for greater trust and deeper relationships between teachers and students. Risk aversion in schools, she argues, has counter-intuitively made it harder for teachers to care and support pupils, in favour of rule enforcement and teaching facts. The book draws on examples from the UK, India, Germany and the US to show how ‘rewilding’ is not just possible, but already happening, in some schools that emphasise education for togetherness, harmony and wellbeing. “Something fundamental needs to change,” Cremin added. “We are crying out for systemic transformation: a completely new vision of what education involves, however challenging that may be.” The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Celebrations at the University of Cambridge honour the life, work and legacy of Sir Robert Edwards, whose work revolutionised fertility treatment through the invention of in vitro fertilisation. Scientists studying human reproduction at the University of Cambridge today are building on Sir Bob Edwards’ incredible legacy. Kathy Niakan A Nobel prize-winner and one of the most influential scientists of the twentieth century, Edwards spent much of his career in the Department of Physiology at the University of Cambridge. Together with gynaecologist Patrick Steptoe and technician and embryologist Jean Purdy, Edwards pioneered the technique of IVF, in which eggs are fertilised by sperm in a laboratory, creating an embryo that is transferred into a woman’s womb to achieve pregnancy. Their breakthrough came when the first IVF baby, Louise Brown, was born in July 1978 - marking the beginning of a new era of medicine. Researchers estimate there have now been over 13 million babies born from IVF worldwide. A two-part event on Friday 26 September at the University of Cambridge will celebrate Edwards’ life, work and legacy, marking what would have been his 100th birthday on Saturday 27 September. An afternoon of talks and discussion, focusing on science and clinical practice, will take part in the Physiology Lecture Theatre - the building where Edwards succeeding in fertilising a human egg in a test tube. It will involve clinicians and scientists who were trained or inspired by Edwards. This will be followed by an evening panel discussion open to the public at Churchill College, Cambridge, where Edwards was a Fellow from 1979 and a Member from 1974. Among the evening panellists will be Louise Brown - the first IVF baby, Dr Jenny Joy - the second of Edwards’ five daughters, Emma Barnett - British Broadcaster and Journalist with a young IVF child, and Dr Mike Macnamee - former CEO of the world’s first IVF clinic, Bourn Hall Clinic, which was established in 1980 by Edwards together with Steptoe and Purdy. “Scientists studying human reproduction at the University of Cambridge today are building on Sir Bob Edwards’ incredible legacy. Many of their careers overlapped with his, and now they’re developing his science further, and also building on his pioneering contributions to the ethics of assisted reproduction,” said Professor Kathy Niakan, Director of the University of Cambridge’s Loke Centre for Trophoblast Research, who will chair the scientific sessions at Friday’s event. She added: “To be part of this field today is a unique opportunity for discovery and innovation, and a great honour to carry forward Sir Bob Edwards’ vision in advancing our understanding of human reproduction.” Dr Jenny Joy, Edwards’ daughter, said, “Our family is delighted to be involved in this event, working with the Loke Centre in the Physiology Department and Churchill College, which both meant a great deal to our father.” Edwards joined the University of Cambridge in 1963, and went on to win the Nobel Prize in 2010 for his work, by which time around four million people had been born following IVF treatment. Edwards died in 2013, aged 87. Infertility affects over 10% of all couples worldwide, and IVF is now one of the most commonly used and successful fertility treatments available. More information about the event is available online. The Bob Edwards centenary conference has been organised by the family of Sir Robert Edwards, the Loke Centre for Trophoblast Research, and Churchill Archives Centre (Churchill College) - which houses Edwards’ papers. The conference is supported by Cambridge Reproduction. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Scientists behind the trial say they are “on the brink of a new class of treatments” and that the findings take us another step closer to stopping disease progression in MS. My instinct is that we are on the brink of a new class of treatments to stop MS progression Dr Nick Cunniffe A combination of metformin, a common diabetes drug, and clemastine, an antihistamine, can help repair myelin – the protective coating around nerves, which gets damaged in multiple sclerosis (MS) causing symptoms like fatigue, pain, spasms and problems with walking. This is according to early findings from the phase two clinical trial, CCMR-Two, carried out by researchers at the University of Cambridge’s Department of Clinical Neurosciences, and funded by the MS Society. The scientists say the results take us another step closer to finally being able to stop disease progression in MS. However, they stress that people should not attempt to acquire the drugs outside a clinical trial, as further research is needed to fully understand their efficacy and safety in MS. Previous evidence from animal studies showed that metformin enhances the effect of clemastine on myelin repair, but until now the two drugs had never been tested together in people. News of the latest trial was presented today at this year’s European Committee for Treatment and Research in Multiple Sclerosis (ECTRIMS) – one of the world’s biggest MS research conferences. “I am increasingly sure that remyelination is part of the solution to stopping progressive disability in MS,” said Dr Nick Cunniffe, a clinical lecturer in Neurology at Cambridge, who led the CCMR-Two trial. “We still need to research the long-term benefits and side effects before people with MS consider taking these drugs. But my instinct is that we are on the brink of a new class of treatments to stop MS progression, and within the next decade we could see the first licensed treatment that repairs myelin and improves the lives of people living with MS.” Over 150,000 people live with MS in the UK. While there are around 20 disease-modifying therapies for people with relapsing MS, and some emerging for active progressive MS, tens of thousands of people remain without effective treatment. Those drugs that do exist only work on one aspect of the condition – the immune system. They don’t stop the gradual nerve damage that leads to long-term disability. Scientists say that protecting nerves from damage by boosting the body's natural ability to put myelin back onto nerves could offer a way forward. “We desperately need ways to protect nerves from damage and repair lost myelin, and this research gives us real hope that myelin repair drugs will be part of the armoury of MS treatments in the future,” said Dr Emma Gray, Director of Research at the MS Society. “These results are truly exciting, and could represent a turning point in the way MS is treated.” Some 70 people with relapsing MS took part in the trials for six months, half of whom took the drug combination and half took a placebo. The primary outcome used to gauge the effectiveness of the drug was a ‘visual evoked potential’ test, which measures how quickly signals travel between the eyes and the brain. The speed of signals slowed down in the placebo group over the course of six months, but remained constant in the drug group. While the primary outcome was positive, scientists point out that people did not feel better on the drugs. The benefit from myelin repair is to insulate and protect damaged nerves, preventing them from degenerating over years. Researchers believe that drugs that promote remyelination will have an effect on disability in the long term, which will be the subject of further research. Researchers argue that MS is just the beginning. Finding ways to protect the brain before irreversible damage sets in is vital across all neurodegenerative conditions, from Alzheimer’s to Parkinson’s. These diseases collectively cost the UK hundreds of billions and place an enormous burden on the NHS and carers. Hannah Threlfell, 43, from Abington was diagnosed with relapsing MS in 2019 after experiencing optic neuritis. She joined the CCMR-Two trial in the hope she could help future generations. “Before I was diagnosed, I sat through a talk from MS specialist Professor Alasdair Coles about groundbreaking MS research. Even though I didn’t know I had it then, I remember thinking how incredible it was that so much had been achieved. And now I have MS, joining the trial was a no brainer,” said Threlfall, a former teacher who has recently become a curate. “I love helping and I know being on this trial will make a difference to someone else in the future – even small ripples have long-lasting effects! This research gives me even more reason to believe that in my lifetime everyone with MS will have treatments that work for them.” CCMR-Two is being funded by donations to the MS Society’s Stop MS Appeal. The appeal hopes to raise £100 million by the end of 2025 to help find treatments that could slow or stop the build-up of disability for everyone with MS. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

As AI becomes embedded in everyday tools and decisions, ensuring the safety and reliability of large language models (LLMs) is more critical than ever. Cambridge spinout Trismik has raised £2.2 million to help it make AI testing faster, smarter and more trustworthy. AI is no longer just generating answers, it's shaping decisions, products and lives. If we want trustworthy AI, we need to treat evaluation as seriously as we take training. Nigel Collier, Professor of Natural Language Processing and co-founder of Trismik LLMs are powering more and more products, but testing their safety, reliability and performance is a significant challenge. Current testing methods are slow, manual and inconsistent, making it difficult for teams to iterate quickly or trust their results. Trismik aims to solve this by using adaptive testing and automatic scoring to evaluate models against a number of dimensions including factual accuracy, bias and toxicity. Inspired by psychometrics and machine learning, the system dynamically selects the most informative test cases, dramatically reducing the number of datapoints required while achieving high reliability and enabling faster development cycles. “AI is no longer just generating answers, it's shaping decisions, products and lives. If we want trustworthy AI, we need to treat evaluation as seriously as we take training. Trismik aims to lead that charge by giving AI engineers the tools to test with precision, act with confidence and build with integrity,” said Nigel Collier, Professor of Natural Language Processing at the University of Cambridge and co-founder and Chief Scientist at Trismik. Collier, who started his career in the 1990s with a PhD in machine translation using neural networks, has increasingly focused on how we can ensure AI acts as a trusted partner to humanity rather than a risk to it. Collier’s curiosity for whether AI could be assessed in the same efficient and fair way as humans, created the genesis for Trismik’s approach to adaptive evaluation. In 2023 Collier met co-founder Rebekka Mikkola, a repeat founder and enterprise sales executive with a passion both for building in AI and opening doors for women in tech. The pair were backed early by Cambridge Enterprise and in 2025 were joined by former Amazon scientist Marco Basaldella as CTO, completing a founding team that blends science, engineering and commercial expertise. Dr Christine Martin, Head of Ventures at Cambridge Enterprise, said: "Trismik exemplifies Cambridge’s continued contribution to global AI development with the team combining world-class academic credentials and practical industry experience that has given them the unique authority to define how AI capabilities should be measured. By solving a pivotal challenge in AI adoption, Trismik is positioned to drive trust at scale - we’re excited to support their journey to market." The £2.2m in pre-seed financing was led by Twinpath Ventures, with participation from Cambridge Enterprise Ventures, Parkwalk Advisors, Fund F, Vento Ventures and angel investors from Ventures Together. Read the full news story on the Cambridge Enterprise website. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Two Cambridge researchers have been named Fellows of the Royal Academy of Engineering in recognition of their exceptional contributions to their fields. Professor Cecilia Mascolo, Professor of Mobile Systems in the Department of Computer Science and Technology and Professor Swami Swaminathan, Professor of Mechanical Engineering in the Department of Engineering, are among 74 leading figures in the field of engineering and technology elected to a Fellowship. This year’s group consists of 60 Fellows, nine International Fellows and five Honorary Fellows. They are drawn from every specialism from within the engineering and technology professions and cover sectors ranging from energy and defence to new materials. They have made exceptional contributions to their field: pioneering new innovations within academia and business, providing expert advice to government, and fostering a wider comprehension of engineering and technology. Professor Cecilia Mascolo, who is also a Fellow of Jesus College, is a pioneer in devising frameworks to collect sensing data from devices such as phones and wearables with the purpose of developing models to understand behaviour and health. During the pandemic, she and her colleagues developed the COVID-19 Sounds App, which collects and analyses short recordings of users coughing and breathing to detect if they are suffering from COVID-19. Since then, she has been working on ways to turn the devices we wear – such as earbuds – into mobile monitors that can collect data about our state of health, and developing cutting edge machine learning tools to evaluate that data on the device itself. Professor Swami Swaminathan, who is also a Fellow of Robinson College, is an expert in the physics and chemistry of turbulent reacting flows, their modelling and simulations. His significant finding in turbulence-scalar-chemistry interaction led to a robust and accurate modelling framework enabling quantitative estimates of temperature distribution, emissions, combustion noise and instabilities in combustors using single simulation. His work helps engineers find robust designs of ‘green combustion systems’ for power generation using low- and zero-carbon fuel and helps devise simple models for complex fundamental phenomena. This year’s new Fellows continue to reflect the Academy’s ongoing Fellowship Fit for the Future initiative announced in July 2020, to drive more nominations of outstanding engineers from underrepresented groups. This commits the Academy to strive for increased representation from women, disabled and LGBTQ+ engineers, those from minority ethnic backgrounds, non-traditional education pathways and emerging industries, and those who have achieved excellence at an earlier career stage than normal. “As we approach our 50th anniversary next year it’s a good time to reflect on how much we have achieved,” said Sir John Lazar CBE FREng, President of the Royal Academy of Engineering. “The Academy is built on the foundation of our Fellowship, and that remains as true today as half a century ago. “Today’s cohort join a community of around 1,700 of some of the most talented engineers and innovators in the UK and around the globe. Their knowledge and experience make them uniquely well placed to tackle the biggest challenges facing the world, and our determination to advance and promote excellence in engineering remains undimmed.” The new Fellows will be formally admitted to the Academy at a ceremony in London on 18 November. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Cambridge scientists are using AI technology to boost research in a range of fields – from better understanding human intelligence, to describing turbulent flows, to freeing computer systems from the cloud – after securing new Fellowships launched to drive breakthrough discoveries. The Encode: AI for Science Fellowships embed top AI talent in the UK’s leading labs to tackle scientific challenges and accelerate the path to real-world solutions. Three Fellowships in the first cohort are being hosted at Cambridge. Encode Fellow Jonathan Carter is using technology originally developed for astrophysics research to decipher how humans understand physics – for example, how the human brain performs intuitive physics calculations, like predicting where a thrown ball will land. Working with Hiranya Peiris, who holds the Cambridge Professorship of Astrophysics (1909), their approach uses interpretable variational encoders, a specialised neural network that can find compact, meaningful representations in complex data. This cross-disciplinary research could advance both our understanding of human intelligence and our ability to build AI systems that learn and generalise like humans do. Shruti Mishra, another Encode Fellow, is developing an AI system that can discover clear, understandable equations describing how turbulent flows behave across different scales. This is a long-standing challenge in physics that affects everything from weather prediction to aerospace design. Guided by Miles Cranmer, Assistant Professor of Data Intensive Science at Cambridge, Shruti is combining machine learning with symbolic mathematics to automatically produce equations that scientists can interpret and trust, rather than ‘black-box predictions’, where the decision-making process is difficult to understand. Their work has the potential to enable more accurate climate predictions and improve industrial designs. And Encode Fellow Martyna Stachaczyk is working with Rika Antonova, Associate Professor at Cambridge, to design a biologically inspired, on-device control architecture for real-time, local intelligence. This research could free intelligent systems from the cloud – which can be insecure and inaccessible where connectivity is limited – enabling robust, adaptive autonomy for prosthetics, robots, and environmental platforms even in resource-constrained or disconnected settings. The Encode AI for Science Fellowship programme is run by Pillar VC, with funding from the Advanced Research + Invention Agency (ARIA) and the UK Government’s Sovereign AI Unit. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

The University has been selected as the lead delivery partner for the British Academy’s new East of England Early Career Researcher Network (ECRN) cluster. Cambridge will work closely with the other delivery partners, Anglia Ruskin University and the University of East Anglia, to support early career researchers in the Arts, Humanities and Social Sciences across the region. Early career researchers are the architects of some of the most innovative and dynamic projects Joanna Page The Academy is completing the national rollout of its ECRN, a researcher-led network for UK-based researchers in Humanities and Social Sciences, as a new cluster is launched to serve the East of England. At Cambridge, the ECRN will be based at CRASSH (the Centre for Research in the Arts, Social Sciences and Humanities) and also supported by the Arts and Humanities Research Facilitation Team and the Postdoc Academy. “Early career researchers are the architects of some of the most innovative and dynamic projects, events, and networks we host at CRASSH, and we are delighted to be able to extend our work with them in this way,” said Professor Joanna Page, Director of CRASSH and academic lead for the East of England Cluster. “The British Academy ECRN will provide a wonderful opportunity for researchers across the region to connect with each other and benefit from a rich programme of research and professional development.” ECRN members benefit from mentoring schemes, training, networking events, grant-writing retreats, academic book-publishing conferences, travel grants to attend network events and conferences, and seed-funding opportunities. “The University of Cambridge has a longstanding commitment to supporting early career researchers, and we are honoured to play a part in this excellent initiative,” said Professor John Aston, Pro-Vice Chancellor for Research at the University of Cambridge. “The British Academy Early Career Researcher Network also helps us to achieve our aims to strengthen ties with academic leaders and communities across the East of England region, helping further the exciting research taking place in Arts, Humanities and Social Sciences.” Daniela Dora, ECR assembly representative for the University of Cambridge School of Arts and Humanities, said: “It is exciting to see the British Academy ECR Network launch in the East of England. The network offers not only new opportunities to share ideas and experiences across disciplines but also provides a supportive community for researchers. For early career researchers, this comes at a crucial stage where collaboration and connection matter most.” The launch event for the East of England cluster of the ECRN will take place on 24 November 2025 in Cambridge, and ECRs from across the region will be invited to take part. Funded by the Wolfson Foundation, Department for Science, Innovation and Technology (DSIT) and Wellcome, the ECRN launched in 2021 as a pilot programme and has since been extended to 2027 due to its success. Find out more and sign up to the ECRN with the British Academy. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

The Artificial Intelligence chatbot, ChatGPT, appeared to improvise ideas and make mistakes like a student in a study that rebooted a 2,400-year-old mathematical challenge. Unlike proofs found in reputable textbooks, students cannot assume that Chat GPT’s proofs are valid Andreas Stylianides The experiment, by two education researchers, asked the chatbot to solve a version of the “doubling the square” problem – a lesson described by Plato in about 385 BCE and, the paper suggests, “perhaps the earliest documented experiment in mathematics education”. The puzzle sparked centuries of debate about whether knowledge is latent within us, waiting to be ‘retrieved’, or something that we ‘generate’ through lived experience and encounters. The new study explored a similar question about ChatGPT’s mathematical ‘knowledge’ – as that can be perceived by its users. The researchers wanted to know whether it would solve Plato’s problem using knowledge it already ‘held’, or by adaptively developing its own solutions. Plato describes Socrates teaching an uneducated boy how to double the area of a square. At first, the boy mistakenly suggests doubling the length of each side, but Socrates eventually leads him to understand that the new square’s sides should be the same length as the diagonal of the original. The researchers put this problem to ChatGPT-4, at first imitating Socrates’ questions, and then deliberately introducing errors, queries and new variants of the problem. Like other Large Language Models (LLMs), ChatGPT is trained on vast collections of text and generates responses by predicting sequences of words learned during its training. The researchers expected it to handle their Ancient Greek maths challenge by regurgitating its pre-existing ‘knowledge’ of Socrates’ famous solution. Instead, however, it seemed to improvise its approach and, at one point, also made a distinctly human-like error. The study was conducted by Dr Nadav Marco, a visiting scholar at the University of Cambridge, and Andreas Stylianides, Professor of Mathematics Education at Cambridge. Marco is permanently based at the Hebrew University and David Yellin College of Education, Jerusalem. While they are cautious about the results, stressing that LLMs do not think like humans or ‘work things out’, Marco did characterise ChatGPT’s behaviour as “learner-like”. “When we face a new problem, our instinct is often to try things out based on our past experience,” Marco said. “In our experiment, ChatGPT seemed to do something similar. Like a learner or scholar, it appeared to come up with its own hypotheses and solutions.” Because ChatGPT is trained on text and not diagrams, it tends to be weaker at the sort of geometrical reasoning that Socrates used in the doubling the square problem. Despite this, Plato’s text is so well known that the researchers expected the chatbot to recognise their questions and reproduce Socrates’ solution. Intriguingly, it failed to do so. Asked to double the square, ChatGPT opted for an algebraic approach that would have been unknown in Plato’s time. It then resisted attempts to get it to make the boy’s mistake and stubbornly stuck to algebra even when the researchers complained about its answer being an approximation. Only when Marco and Stylianides told it they were disappointed that, for all its training, it could not provide an “elegant and exact” answer, did the Chat produce the geometrical alternative. Despite this, ChatGPT demonstrated full knowledge of Plato’s work when asked about it. “If it had only been recalling from memory, it would almost certainly have referenced the classical solution of building a new square on the original square’s diagonal straight away,” Stylianides said. “Instead, it seemed to take its own approach.” The researchers also posed a variant of Plato’s problem, asking ChatGPT to double the area of a rectangle while retaining its proportions. Even though it was now aware of their preference for geometry, the Chat stubbornly stuck to algebra. When pressed, it then mistakenly claimed that, because the diagonal of a rectangle cannot be used to double its size, a geometrical solution was unavailable. The point about the diagonal is true, but a different geometrical solution does exist. Marco suggested that the chance that this false claim came from the chatbot’s knowledge base was “vanishingly small”. Instead, the Chat appeared to be improvising its responses based on their previous discussion about the square. Finally, Marco and Stylianides asked it to double the size of a triangle. The Chat reverted to algebra yet again – but after more prompting did come up with a correct geometrical answer. The researchers stress the importance of not over-interpreting these results, since they could not scientifically observe the Chat’s coding. From the perspective of their digital experience as users, however, what emerged at that surface level was a blend of data retrieval and on-the-fly reasoning. They liken this behaviour to the educational concept of a “zone of proximal development” (ZPD) – the gap between what a learner already knows, and what they might eventually know with support and guidance. Perhaps, they argue, Generative AI has a metaphorical “Chat’s ZPD”: in some cases, it will not be able to solve problems immediately but could do so with prompting. The authors suggest that working with the Chat in its ZPD can help turn its limitations into opportunities for learning. By prompting, questioning, and testing its responses, students will not only navigate the Chat’s boundaries but also develop the critical skills of proof evaluation and reasoning that lie at the heart of mathematical thinking. “Unlike proofs found in reputable textbooks, students cannot assume that Chat GPT’s proofs are valid. Understanding and evaluating AI-generated proofs are emerging as key skills that need to be embedded in the mathematics curriculum,” Stylianides said. “These are core skills we want students to master, but it means using prompts like, ‘I want us to explore this problem together,’ not, ‘Tell me the answer,’” Marco added. The research is published in the International Journal of Mathematical Education in Science and Technology. The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.