Latest AI (Artificial Intelligence) News
OpenAI Predicts 2026 as Breakthrough Year for AI in Scientific Research
OpenAI's Kevin Weil forecasts 2026 as pivotal for AI reshaping science, with GPT-5.2 scoring 92% on GPQA benchmark, surpassing human experts at 70%. Scientists not using AI will face disadvantages; models now act as humble brainstorming partners with epistemological humility.
Focus shifts to accelerating discoveries by digesting past papers.
Chinese Researchers Achieve AGI Breakthrough in Logical Reasoning with TongGeometry
BIGAI and Peking University team developed TongGeometry, outperforming DeepMind's AlphaGeometry in geometry problem-solving and autonomous generation. System uses normalized representation to solve path explosion, enabling 'small data, big task' paradigm mimicking human intuition.
Three AI-generated problems selected for 2024 Chinese Mathematical Olympiad; supports automated proofs and intelligent education.
OpenAI Launches Prism AI Workspace for Scientific Research
Prism, powered by GPT-5.2, integrates into research workflows as a cloud-based environment for hypothesis testing and analysis. Part of OpenAI's push toward automated AI research interns by September 2026.
Aims to accelerate biology, chemistry, and materials science discoveries.
UK Advances AI Opportunities Action Plan with Supercomputers and Funding
UK launched Isambard-AI supercomputer and plans sixfold capacity increase at Cambridge; AI Safety Institute tested 30 frontier models. Sovereign AI Unit to launch April 2026 with ÂŁ500M for UK AI firms.
AI Assurance Innovation Fund opens Spring 2026.
Agentic and Physical AI Set for Enterprise Adoption in 2026
Agentic AI expected to represent 10-15% of IT spending in 2026, with 33% of enterprise apps by 2028; physical AI adoption to hit 80% in manufacturing. Vision-Language-Action models like NVIDIA's GR00T drive robotics and autonomous vehicles.
Humanoid robotics funding exceeded $2B in 2025.