The EU is investigating OpenAI and Anthropic following incidents involving AI agents - Computerworld