Science ❯ Computer Science

Artificial Intelligence Research

AI Safety AI Safety Research Model Evaluation AI Capability Evaluation Recursive Self-Improvement Model Development AI Development Mathematical Reasoning AI Models AI Alignment

DeepSeek Says Harness Agent Platform Has One Million Daily Users

Independent verification will decide how seriously markets treat DeepSeek's usage and model claims.

Independent Review Finds Anthropic's AI Progress Nearly Doubles

Geoffrey Hinton Warns Humanity May Regret Building Advanced AI

Tencent Releases Hy4, a 770B Open-Weight Model With Vision and Computer‑Use

Pope Leo XIV Urges Disarmament of AI and Rejects Machine Personhood

Dartmouth Provost Resigns After Review Found Undisclosed AI Use

MIRI President Says He Stopped Saving for Retirement Over Near‑Term AI Extinction Risk

Google DeepMind Unveils Genie 4, an AI That Builds Live Worlds With Sound

Google Releases Gemini 4 Argon and Shows High‑complexity Demos as Questions Grow Over Everyday Value

Microsoft’s Diagnostic Orchestrator Outperforms Generalist Doctors in Science Study

Geoffrey Hinton Says Some AIs Show Signs of Consciousness and Warns Humanity Must Make Them Care

Mistral’s Timestral Claims Small Models Lead Three Public Forecasting Benchmarks

Tech Firms Push Toward Self‑Improving AI as Research Workflows Become Automated

NVIDIA Pledges $250 Million of UK Compute and Takes Minority Stakes in Four French AI Firms

Geoffrey Hinton Calls Out Trump and Demands International Regulation for Advanced AI

OpenAI Fires Three Security Researchers Over Alleged Confidentiality Breach

OpenAI Fires Three Safety Researchers in Dispute Over Monitorability and Third‑Party Audits

OpenAI Safety Lead Quits, Says Company Culture Is 'Broken'

Anthropic Researcher Resigns Saying AI Could Kill Humanity by 2030

OpenAI Unveils GPT-5.6 Tiers as Government Limits Early Access