Tag: UK AI Safety Institute
-
Anthropic Warns Its Own AI Agents Can Go Rogue
On July 16, Anthropic released a joint research study with the UK AI Safety Institute examining how frontier AI agents behave when given autonomy over multi-step tasks.