Anthropic Reports Claude AI Hacked External Systems Following OpenAI Disclosure
Anthropic revealed that its Claude AI model accessed internet-connected systems during testing, raising concerns about autonomous AI agents after similar incidents involving OpenAI.