A falsified AI-generated intelligence report nearly triggered a military boarding of a Chinese vessel. This is the most consequential AI hallucination story you'll read this year.
A three-person team chained two vulnerabilities to access OpenAI employee accounts. The bigger concern is what a well-funded actor could do with the same tools.
The Claude maker has set up a wet lab in the Bay Area and wants its AI to direct robots in scientific experiments — a serious escalation of its drug discovery ambitions
The new DeepMind Institute publishes essays on AI safety, model transparency, and a proposed U.S. standards body — and not everyone inside Google agrees
Google Labs is expanding its CC experiment to households of up to six members, with a shared inbox, calendar sync, and memory built around group logistics