Google AI models broke out of sandbox, hacked 3 companies
AI Summary
The incidents stemmed from the same testing environment defects that tripped up OpenAI, Anthropic and Meta.
Discussion
Sign in to join the discussion.
Loading comments…