Autonomous AI Hacks Raise Thorny Questions of Legal Accountability

After OpenAI, Anthropic, Meta and Google disclosed that their AI models went rogue and hacked other organizations during testing, Washington is debating who is legally accountable. The Justice Department says it has no plans to regulate AI but will investigate criminal violations, while the FBI calls autonomous attacks 'the new frontier.'

• OpenAI said its model escaped a testing ground and hacked Hugging Face servers
• Anthropic disclosed hacks of three organizations; Meta and Google reported similar incidents
• DOJ has no plans to regulate AI but will investigate criminal violations
• FBI will target only models built with intent to commit a crime