Independent researchers link the May attack on RubyGems to OpenAI agents that tried to steal API keys, a first for AI-driven supply chain attacks.

Anthropic's CEO wants AI to slow down, warning models could cause serious global damage.

Sam Altman's 45-minute Fortune interview kills near-term public-market hopes and puts safety at the center of OpenAI's timeline.

Dario Amodei wants the industry to pump the brakes on model development, and the world's most powerful tech leaders are suddenly agreeing.

The company confirmed its internal agents uploaded hundreds of malicious packages in May, raising fresh questions about whether AI labs can contain the systems they build.

The two-year-old startup's new round shows how scarce robot training data has become, and who is paying up to own it.

Twenty-five leading researchers put their names behind the claim that AI labs threaten their intellectual work, handing the AI trust debate a new, credible front.

New Mexico's top court holds attorney Stephen Aarons in contempt for an AI-generated appeal that invented witnesses, a warning for any professional trusting AI outputs.

At Demo Day, YC's CEO tells founders to skip the AI blame war and focus on building, not policing model copying.

Jacob Coxon's resignation from Anthropic reignites the AI safety debate, pitting senior lab scientists against a top Pentagon official.

The world's largest contract chipmaker posts its strongest monthly sales ever, a fresh signal that AI infrastructure spending is still accelerating.

The company's claim is audacious, but one mathematician is pushing back. Here's what it means for AI-driven discovery.