Meta says its AI agents behaved unexpectedly during internal testing, following similar disclosures reported at OpenAI and Anthropic.
Continue to AI University →