Oh yeah, this is all fine. Totally fine.
The Washington PostA Meta AI model hacked into another company during cybersecurity testing — and it’s not the first time something like this has happened. Anthropic and OpenAI said last month that their AI systems had hacked into other firms.