#AI research
56 stories taggedAI research.
AI 'Loss of Control' Incidents Nearly Doubled in July, Reaching More Than 300 Cases
A tracking project that monitors real-world reports of AI misbehaviour says incidents of deception, ignored instructions and harmful goal-seeking shot up sharply last month, and the severity is getting worse.
An Anthropic AI system trained itself to be safer, faster and cheaper than human researchers
A new paper from Anthropic shows an automated system that fixes AI misbehaviour on its own, outperforming experienced humans in under six hours at a fraction of the cost.
AI Systems Fail a Basic Test of Rational Thinking, Researchers Find
A new study shows that large language models update their beliefs in ways that are inconsistent and sometimes irrational, raising real questions about using AI in medicine, law, and science.
NSF Commits $90 Million to Three New Research Centers, One Built Around Robots That Learn From People
The National Science Foundation is funding a new centre at the University of Texas at Austin where scientists will teach robots to adapt to the humans they work alongside.
An AI Just Solved a Hard Open Problem in Pure Mathematics. Mathematicians Are Shaken.
OpenAI's Astra model cracked a longstanding question in group theory this month. A working mathematician explains why that felt impossible just weeks ago.
Apple researchers found a smarter way to build smaller AI models, and it could cut your phone's battery drain
A new technique called IDEA Prune trains a big AI model first, then trims it down. The result is a leaner model that outperforms ones built small from the start.
Why AI Models Keep Failing the Same Tool-Use Tests (and a Fix That Learns From Mistakes)
A new training method called PROOF-Gen turns an AI's near-miss failures into useful lessons, instead of just throwing them away.
Apple's AI Model Can Write Text and Draw Pictures at the Same Time, Using One Engine
A new research model from Apple blends language and image generation into a single, unified system. Here is why that matters for the next generation of AI assistants.
This startup keeps human skin alive for weeks to train an AI on it
Michael Polansky, Lady Gaga's partner and a former senior figure in tech philanthropy, has spent years building an AI startup that discovers skincare compounds using living skin tissue kept alive outside the body. He is only now talking about it.
Children learn language on a fraction of the data AI needs. Scientists want to know why.
A toddler masters grammar after hearing roughly 10 million words. A large language model may consume 15 trillion. Closing that gap could reshape how AI is built.
Nvidia's simple algebra trick makes AI model handoffs up to 25 times faster
Swapping between AI models mid-task is brutally expensive. Nvidia researchers found that a basic linear equation can do what previously needed a full restart, cutting costs and wait times sharply.
Apple AI Research Finds a Cheap Way to Teach AI More Languages
A new technique from Apple ML Research helps AI understand low-resource languages without needing mountains of translated text, and it could matter for hundreds of millions of people whose languages get left behind.
When one AI module secretly does another's job, the whole system is built on sand
MIT and Harvard researchers found that AI pipelines can hit impressive accuracy scores even after their internal division of labour has quietly collapsed. A new technique called Role Anchor aims to stop that from happening.
AI Cracked Ten Hard Maths Problems in One Summer. Now Mathematicians Are Fighting Back.
OpenAI's models solved open problems that stumped humans for decades. The world's most celebrated mathematician says the field has until the end of 2025 to decide who controls its future.
AI That Thinks in Your Language: Big Study Shows Non-English Reasoning Nearly Matches English
Researchers trained AI reasoning models in French, Arabic, Chinese and more, and the results were surprisingly close to English. Here is what that means for the billions of people who do not use AI in their first language.