OpenAI Released 719 AI Maths Manuscripts: What Counts as a Verified Result?
OpenAI’s mathematics release separates formal proof checking from human understanding, scholarly significance and community verification.
Category: AI Research
Can an AI Model Really Forget? Machine Unlearning Is Not a Delete Button
Machine unlearning promises to remove unwanted influence from AI models, but recent research shows why refusal, deletion and verified forgetting are not the same.
AI Is Making Science Faster—But Verification Is Becoming the Bottleneck
New evidence suggests AI can save researchers time and generate more hypotheses, while shifting the hard work downstream into validation, experiments and judgement.
An AI Agent Demo Is Not a Reliability Test: How to Read Long-Horizon Benchmarks
Long-horizon AI-agent benchmarks reveal what polished demos hide: repeated reliability, partial progress, recovery, cost and efficiency across complex multi-stage work.
When AI Trains on AI: Model Collapse and the Case for Preserving Human Data
Model collapse can make generative AI narrower and less reliable when synthetic outputs feed future training. Learn what the evidence shows and how to protect original data.
A Successful AI Agent Can Still Be Unsafe: Test Permission, Recovery and Human Handoffs
A practical framework for evaluating AI agents across task accuracy, permission discipline, traceability, human escalation, recovery and operating cost.
The Model Passed Its Test. Production Is a Different Experiment
An AI system can pass laboratory tests and still fail in production. Build a monitoring plan for drift, uneven errors, incidents, human overrides and retirement.
AI Is Not Just Helping Students Learn—It Is Changing Where Thinking Happens
A new Communications Psychology perspective argues that generative AI should not be treated as another classroom tool. Because AI can explain, evaluate, generate and interact continuously, it may redistribute the cognitive work traditionally performed by students and teachers.
Texas Freezes New Data-Centre Permits: AI’s Next Bottleneck May Be Physical Infrastructure
Texas has halted new state permits for data centres while officials audit their impact on electricity, water and infrastructure as AI demand accelerates.
OpenAI Creates Independent Mathematics Advisory Group as AI Moves Deeper Into Scientific Discovery
OpenAI has created an independent advisory group for mathematics and AI as advanced models move further into scientific reasoning and mathematical discovery.