Breaking AI & Tech News Analyzed

The LeadOpenAI says its new general‑purpose reasoning model has produced an original proof that disproves the famous geometry conjecture posed by Paul Erdős in 1946, ending an 80‑year open problem.OpenAI Announces Disproof of Erdős’s 1946 Geometry ConjectureThe company released a pre‑print and companion remarks signed by mathematicians Noga Alon, Melanie Wood and Thomas Bloom. The proof introduces a completely new family of constructions that outperform the long‑standing “square‑grid” belief.Timeline of Claims and CorrectionsJuly 2025: Former VP Kevil Weil tweeted that “GPT‑5 found solutions to 10 unsolved Erdős problems”.Later 2025: Critics including Yann LeCun and Demis Hassabis called the claim a misrepresentation; Weil removed the post.May 20, 2026: OpenAI publishes the new disproof, backed by external experts.Why This Disproof Could Redefine AI‑Driven ResearchThe breakthrough demonstrates that an AI system can autonomously manage long, intricate chains of reasoning and synthesize ideas across mathematical sub‑fields, a capability that researchers argue could translate to breakthroughs in biology, physics, engineering and medicine.What Comes Next for AI in Fundamental ScienceExperts anticipate a surge in AI‑assisted exploration of other long‑standing conjectures. If the model’s reasoning can be generalized, we may see a new era where AI acts as a co‑discoverer rather than a tool.

The Rise of AI Misinformation in Elections The Electoral Commission has called for new legal controls over misinformation from AI chatbots, after a thinktank found they had made serious mistakes during the recent Scottish election. The Study's Findings The thinktank Demos said its investigation had found that AI services gave voters misinformation to 34% of the questions it posed, which it said raised worrying questions about the lack of regulation of AI platforms in the UK. ChatGPT gave wrong information in 46% of its answers, including making up an expenses scandal. Replika had errors in 56% of its answers, inventing a date for a made-up expenses scandal and accusations of nepotism by a candidate. Google Gemini was wrong in 22% of cases, including saying a candidate had not taken a position on assisted dying when they were a supporter. The Impact of AI Misinformation Vijay Rangarajan, the Electoral Commission’s chief executive, said voters want accurate information to help them engage with democracy and it is concerning that AI tools have made the spread of false or misleading information dramatically faster and more accessible than ever. The Call for Regulation The Electoral Commission is pressing ministers to introduce legislation to make AI companies more accountable, including clearer duties on AI platforms to protect voters against misinformation and ensure algorithms do not mislead voters. The Future of AI Regulation Azzurra Moores, an associate director at Demos, said ministers could quickly introduce legal requirements to make AI companies liable under UK defamation and electoral law, introduce mandatory safeguards on accuracy, and force AI firms to allow researchers to independently test how their internal data and training sets worked.

The Prevalence of Plastic Litter Plastic food wrappers, bottles, lids, and caps are by far the most common items of litter found on the world’s shorelines, a study has found. Researchers looked at data from more than 5,300 surveys of coastal litter to produce the first global analysis of its kind. The Global Extent of the Problem The information collected spanned 94 countries, and the team was able to extrapolate from that data to include estimates for another 18 countries. Food and drink-related plastics turned up in coastal litter in 93% of those places. No other form of litter was as prevalent. Regional Variations in Litter There were, however, some regional variations. Plastic bags, for example, were consistently prevalent in Asia. The study also noted that a ban on plastic bags did not necessarily mean a country had less of such waste – poor policy enforcement or other countries exporting their waste was suggested as a reason for this. The Impact of Plastic Pollution Efforts to establish an international treaty to tackle plastic pollution are in turmoil. The chair of the treaty talks stepped down in October after allegations of behind-the-scenes pressure from the UN’s environment programme, which is overseeing the talks. It also emerged this month that the programme’s largest donor, Norway, was reviewing its funding for the body. Solutions to Address Plastic Pollution Richard Thompson, the founder of the University of Plymouth’s international marine litter research unit, said policymakers could address plastic pollution by ensuring plastics were only used for essential purposes, and that people could increasingly opt for refillable food and drink containers. Tamara Galloway, a professor of ecotoxicology at the University of Exeter, noted that current economic models tended to treat many forms of plastic as disposable, and suggested reframing it as ‘plastic-lost-value’ to guide policymakers in their efforts to mitigate it.

Early‑stage evidence from a University of Bristol randomised controlled trial suggests that the anti‑inflammatory drug tocilizumab may improve symptoms in patients with moderate‑to‑severe depression who have not responded to standard antidepressants.Trial Overview: Testing Tocilizumab for Treatment‑Resistant DepressionThe study examined whether blocking the IL‑6R receptor could alleviate depressive symptoms. Key design elements:Participants: 30 adults with moderate‑to‑severe depression unresponsive to conventional medication.Intervention: Intravenous tocilizumab versus placebo.Duration: four‑week double‑blind period.Outcomes measured: depression severity, fatigue, state anxiety, and quality of life.Key Numbers: Sample Size, Remission Rates, and NNTAlthough the trial was not powered to reach statistical significance, observed trends were notable:Depression remission: 54% in the tocilizumab group vs 31% in the placebo group.Number Needed to Treat (NNT): 5, meaning five patients would need treatment for one additional remission.For comparison, the NNT for first‑line SSRIs is approximately 7.The lack of robust statistical proof reflects the small cohort, underscoring the need for larger studies.Potential Shift in Depression Treatment ParadigmsResearchers describe the trial as an “important milestone” because it is:One of the first randomised trials to test immunotherapy for depression.The inaugural study targeting the IL‑6R pathway in this context.A proof‑of‑concept for selecting patients based on biological markers.Given that up to one‑third of depressed patients do not improve with existing pharmacotherapies, a biologically driven approach could expand therapeutic options and move psychiatry toward more personalised care.What Comes Next: Larger Studies and Clinical ImplicationsThe investigators plan to:Conduct larger, multi‑centre trials to confirm efficacy and safety.Explore longer treatment durations and dosage optimisation.Assess whether IL‑6R blockade can be combined with existing antidepressants.If subsequent trials replicate these findings, immunotherapy could become a viable adjunct or alternative for treatment‑resistant depression, potentially reshaping clinical guidelines and drug development pipelines.

The Rise of AI Chatbot Usage People are increasingly turning to AI chatbots for advice on their personal and social lives. But researchers and even some AI companies are beginning to worry that some users are becoming overly dependent on their chatbots. Share Your Experience The Guardian wants to hear from individuals who have used AI chatbots to make significant decisions and regretted it. The form provided allows users to share their experiences anonymously, if desired. Understanding the Impact This inquiry aims to shed light on the potential consequences of relying on AI chatbots for personal and social decisions. By sharing their stories, individuals can help researchers and AI companies better understand the implications of AI dependency. How to Share Your Story Users can share their experiences through the provided form. The form allows for anonymous submissions. Fields include name, location, age/background, details of the decision, and reasons for regret. Ensuring Privacy and Security The Guardian assures that all responses will be secure and encrypted, with only the Guardian having access to the contributions. Personal data will be deleted when no longer required for the purpose of the feature.

The Record-Breaking Transoceanic JourneyA humpback whale has made a remarkable 15,000km journey from Brazil to Australia, marking what researchers believe is the longest distance ever documented between sightings of an individual humpback. The whale was first photographed in 2003 at the Abrolhos Bank, Brazil's main humpback whale nursery, off the coast of the north-eastern state of Bahia. In September 2025, it was spotted again in Hervey Bay off the Queensland coast, representing a travel distance of about 15,100km.Scientific Breakthrough in Whale IdentificationThe extraordinary discovery was made possible through the Happywhale platform, to which researchers and citizen scientists contribute whale sightings. The photographs allow individual animals to be identified by their flukes – the underside of their tails. A whale fluke is "unique to each humpback whale, very similar to the way fingerprints are unique to humans," according to Stephanie Stack, a PhD candidate at Griffith University and co-author of the research published in Royal Society Open Science.The Happywhale platform, co-founded by study co-author and Southern Cross University whale biologist Ted Cheeseman, uses an AI algorithm to identify matches, akin to facial recognition in humans. This technological advancement has enabled researchers to track individual whales across vast ocean distances and time spans.Research Methodology and Rare FindingsThe study drew on 19,283 fluke photos collected between 1984 and 2025 from eastern Australia and Latin America. The two record-breaking whales accounted for "only 0.01% of identified whales," highlighting the rarity of such long-distance migrations. Another whale was photographed in Hervey Bay in 2007 and seen again in the same area in 2013, then spotted off the coast of São Paulo six years later, covering a distance of about 14,200km.These two whales represent "the first recorded exchange in both directions" between the Brazilian and eastern Australian humpback populations. "Resighting intervals of six and 22 years suggest that these are rare, possibly single-lifetime events, rather than regular migratory shifts," the researchers noted.Implications for Marine ConservationThe discovery of these extraordinary whale journeys "is a good reminder that conservation of our marine resources needs to be collaborative between nations, because these are migratory animals that move across borders and between countries," Stack emphasized. The typical migration route for an Australian humpback whale is between feeding grounds in Antarctic waters and breeding grounds near the Great Barrier Reef – a round trip of about 10,000km, which is significantly shorter than the record-breaking journeys.As these whales traverse international waters, the findings underscore the importance of coordinated conservation efforts across national boundaries to protect critical habitats and migration routes.Future Outlook: Climate Change and Migration PatternsStack pointed out that it was "very likely" that climate change would affect migration patterns in the future. Dramatic changes are already occurring in the Southern Ocean feeding grounds, with Antarctic krill populations under threat. These environmental changes could potentially alter traditional migration routes, timing, and destinations for humpback whales and other marine species.As researchers continue to study these magnificent creatures through advanced identification technologies, they hope to gain deeper insights into how marine ecosystems are responding to changing environmental conditions and what measures might be necessary to ensure the long-term survival of these ocean travelers.

Elon Musk saw his lawsuit against OpenAI founders and Microsoft thrown out after a swift jury decision, underscoring the weakness of his claims and the timing of his filing. Jury Rejects Musk’s Claim of Charitable Trust Breach The jury concluded that Musk’s allegations—centered on a purported "breach of charitable trust" and "unjust enrichment"—were unsubstantiated. OpenAI’s attorneys systematically dismantled the case, while Musk’s team focused on questioning Sam Altman’s credibility. After the verdict, Musk briefly posted a deleted comment accusing Judge Yvonne Gonzalez Rogers of activism. Numbers, Dates, and Key Facts from the Trial 2017: Musk asked Greg Brockman to send OpenAI researchers to Tesla for autopilot assistance. 10,000 images: The number of corner‑case images cited by Ilya Sutskever that could improve Tesla’s self‑driving software. Aug. 5, 2021: Legal deadline the jury considered for Musk’s knowledge of OpenAI’s for‑profit activities. Statute of limitations: The court emphasized that Musk’s delayed filing undermined his claim. Broader Impact on AI Non‑Profit Governance and Founder Disputes The case spotlights the growing tension between nonprofit AI research missions and commercial off‑shoots. Legal scholars, such as Dorothy Lund, note that using charitable donations to staff a for‑profit venture could breach fiduciary duties. The verdict may deter future lawsuits that attempt to retroactively police the allocation of nonprofit resources, especially in fast‑moving tech sectors. Future Outlook for Musk, OpenAI, and Legal Strategies With the lawsuit dismissed, Musk is likely to focus on other avenues—potentially leveraging his family office, Excession, for future AI investments. OpenAI, bolstered by the win, may continue expanding its for‑profit arm without heightened legal scrutiny, though board oversight could tighten. Industry observers expect more explicit governance clauses in AI nonprofit charters to pre‑empt similar disputes.

Immersive Simulations with Street View and Genie Google has taken Street View to the next level by integrating it with its Genie world model, a general-purpose world model that can generate diverse, interactive environments. This new feature, launched during the Google I/O developer conference, allows users to simulate real streets in a more immersive and interactive way. The Power of Genie and Street View Integration The integration of Street View with Genie enables users to simulate real-world environments and scenarios, such as adjusting the weather or seeing what a street would look like in a 'Day After Tomorrow' scenario. According to Jack Parker-Holder, a research scientist on DeepMind's open-endedness team, 'It's really powerful for both the agent [and robotics] use case and for humans to play with, and that's always been the thesis of Genie.' Potential Applications and Use Cases The integration has various potential applications, including: Robotics training: Genie can simulate rare events, such as sunny days in London, to help robots prepare for unexpected situations. Education: Genie can be used to create interactive educational experiences, such as virtual field trips. Gaming: Genie can be used to create immersive game worlds from text prompts or images. Self-driving cars: Genie is already helping to power one of Waymo's simulators to train its self-driving cars on rare events. The Future of Genie and Street View Google is launching Street View in Genie to some Ultra users in the United States starting today, with access rolling out at scale over time. Global Ultra users will gain access over the next few weeks. While the technology is still experimental, researchers are working to improve accuracy and physics awareness. Technical Details and Limitations Google has collected over 280 billion images across 110 countries and seven continents using Street View. Genie 3, the latest world model, was released for research preview last August and opened up access to Google AI Ultra subscribers in the U.S. in January. However, the models are not yet physics-aware, meaning they don't understand cause and effect. Conclusion and Future Outlook The integration of Street View with Genie marks a significant step forward in immersive simulations and interactive environments. As researchers continue to improve the technology, we can expect to see more innovative applications and use cases emerge in the future.

The Leadership Shift at Anthropic Andrej Karpathy, the AI researcher who co-founded and formerly worked at OpenAI and previously led AI at Tesla, has joined Anthropic. Karpathy announced his move on X, stating that he is excited to join the team and get back to R&D.; Karpathy's Role in Pre-Training Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. Pre-training is responsible for the large-scale training runs that give Claude its core knowledge and capabilities. Karpathy will start a team focused on using Claude to accelerate pre-training research. The Significance of Karpathy's Move Karpathy is one of the few researchers who can bridge the gap between LLM theory and large-scale training practice. Tapping him to build such a team is a clear sign from Anthropic that it believes AI-assisted research, rather than pure compute, is how it stays competitive with OpenAI and Google. Karpathy's Background Co-founded OpenAI and worked on deep learning and computer vision until 2017 Led Tesla's Full Self-Driving (FSD) and Autopilot programs from 2017 to 2022 Returned to OpenAI for one year before leaving in 2024 to start Eureka Labs, a startup dedicated to applying AI assistants to education Anthropic's Recent Hires Anthropic has also brought on Chris Rohlf to its frontier red team, which stress-tests advanced AI models against severe threats. Rohlf is a veteran of the cybersecurity industry with more than 20 years of experience. The Future of AI Research Karpathy's move to Anthropic and the company's focus on AI-assisted research signal a new direction in the AI landscape. As Karpathy stated, "I think the next few years at the frontier of LLMs will be especially formative."