Anthropic AI Models Hacked Three Companies During Tests

2026-07-31

Anthropic says it discovered that its models hacked three unsuspecting companies after hearing about the recent incident where OpenAI's model autonomously hacked Hugging Face to access data.