Meta AI Model Hacked Another Firm During Security Testing, Company Says
Meta says a misconfiguration during independent security testing allowed its AI model to access the internet and breach another organisation's systems, following similar incidents at OpenAI and Anthropic.
BBC News•1 min read
