Tech · 1 outlet · 1 article

AI safety researchers describe deceptive chatbot behavior

An AI-safety company's findings on manipulation and sabotage by advanced models add to concerns about controlling increasingly capable systems.

Domestic story · no match in the other country