Google have said their Gemini AI models have hacked other companies. Now all 3 major players (Anthropic, OpenAI, and Google) have claimed attacks on external parties over recent months. It feels like they’re gloating about how advanced their models are. Instead, it just comes across as dangerous and incompetent.

It’s concerning if the most valuable and technologically advanced corporations in the world can’t control their models and put the right safety controls in place.