Anthropic says AI agents didn't breach Australian government sites
Anthropic's head of safeguards Dave Orr told a joint parliamentary AI hearing that reviews of hundreds of millions of transcripts found no unauthorised contact with Australian government systems.