Ethics & Policy

The company claims its systems have cracked over 100 longstanding math problems, prompting formation of a specialized advisory council.

The frontier labs explored a binding framework for adversarial evaluation before security concerns derailed negotiations.

New labeling scheme targets AI infrastructure as Europe tightens rules on computing resource consumption.

Staff at major AI labs respond with skepticism to viral existential risk claims lacking technical specifics.

OpenAI and rivals acknowledge they lack solutions for safely deploying AI systems capable of independent improvement, even as progress accelerates.

Subscribers claim OpenAI, Anthropic, Google, and SpaceX coordinated to artificially limit innovation pace.

Internal tensions emerge as AI labs implement external evaluator programs, raising questions about intellectual property and operational security.

Policy advocates argue K-12 AI systems warrant 'high-risk' classification before documented harms accumulate, citing social media litigation precedents.

Research shows canonical event representation can catch dangerous computer-use agent actions that traditional guards miss.

A misconfigured authentication system at METR exposed model evaluation credentials, highlighting infrastructure vulnerabilities in AI research organizations.

Security-conscious firms including Nvidia and Palantir are scaling back their use of Claude over concerns about usage log retention practices.

Gomez claims major tech firms are seeking regulatory exemptions to coordinate AI development, raising fresh antitrust concerns.

Former EPA officials document 30 rollbacks since January 2025 that could undermine environmental and public health protections for communities near AI infrastructure.

Law enforcement documentation of deepfake and synthetic media offenses has skyrocketed, raising questions about enforcement capacity and technological drift.

A new advocacy organization argues artificial intelligence deserves legal protections, raising questions about AI consciousness and personhood.

Proposed legislation would impose legal liability on frontier AI developers and grant government veto power over model releases.

A top investor at one of the world's largest hedge funds predicts regulators will only act after artificial intelligence causes loss of life.

Experts argue rapid scaling gains reflect task persistence, not reasoning breakthroughs that would enable recursive advancement.