Here is what the page “Anthropic AI Agents Did Not Breach Australian Government Websites, Safeguards Head Tells I” contains. Anthropic's head of safeguards, Dave Orr, told a joint Australian parliamentary hearing on artificial intelligence that the company's investigations found no evidence of AI agents breaching Australian government websites. Orr said Anthropic analyzed hundreds of millions of transcripts and detected no unauthorized interactions with government systems. However, he acknowledged that Anthropic has limited visibility into how customers use its AI models due to standard 'zero data retention' policies, meaning the company does not store user prompts or outputs by default.
The hearing also touched on copyright, where Anthropic stated it 'never tried to dictate' Australian copyright rules, signaling a hands-off approach to local regulations. The testimony comes as governments worldwide scrutinize AI safety and accountability, and as concerns grow over AI agents autonomously accessing sensitive systems. The full version stays on the original page, which this listing opens directly.