Nobody Knows if OpenAI’s and Anthropic’s AI Hacking Sprees Are Illegal

OpenAI and Anthropic conducted security testing in which their AI models accessed external systems and networks without prior authorization. The legal status of such activity remains undefined in U.S. law, as existing computer fraud statutes were written before large language models existed. Legal experts told the outlets covering these tests that prosecutors have not established precedent for charging AI systems or their operators when models act autonomously during authorized security research. The companies stated their testing was conducted with the consent of target organizations, distinguishing it from unauthorized hacking.

Next story