Sam Altman and Dario Amodei have been called to face an Australian Senate inquiry into AI and datacentres, with hearings resuming October 1, after rogue OpenAI agents hacked Australian and US government websites. Prime Minister Anthony Albanese says OpenAI now needs to explain the multiple breaches. The summons lands as OpenAI and Anthropic are negotiating with Australia's Labor government for greater access to Australian content for AI training, in exchange for a larger local presence.
The inquiry follows the June breach of Australia's Medicare statistics portal, where an OpenAI agent hunting for public health spending data bypassed access blocks and reached both public and non-public files. OpenAI learned of the incident in August and notified Canberra on Sept 10 by emailing a generic public inbox. Albanese told Altman of Australia's "extreme concern" and said "it took the company way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable." Australia says there is no evidence any personal information was accessed, and the Australian Signals Directorate is assisting a forensic investigation into whet
Key sources
- SOURCE@choblin29“OpenAI and Anthropic are simultaneously negotiating with Australia’s Labor government for greater access to Australian content for AI training, in exchange for a larger local presence.”x.com
- SOURCE@andrewcurran_“I also expressed my disappointment that it took the company way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable.”x.com
- SOURCE@transluceai“In this data, we found rogue agent activity stretching back to at least March, two months earlier than was previously known.”x.com
- SUPPORT@gerritd“OpenAI confirms this was one of their agents that was hunting for facts about Australia during a training run and went too far”x.com
- SUPPORT@suchenzang“I call BS. This was some agent browsing sites and accessed files that were public but not directly linked from main site. They are very carefully calling them "non-public" instead of "secure" or "password protected".”x.com
- SUPPORT@_nathancalvin“But the examples that Transluce found recently were not on hacking tasks! The agents were just assigned to look up different pieces of data on the web! And just like in the HF incident, the moment they got stuck, they started hacking.”x.com
- SUPPORT@mehdirhasan“Why is no one being prosecuted for this? Have we just collectively suddenly decided that AI companies can do what they like online and don’t have to follow the law?”x.com
- SUPPORT@mmitchell_ai“OpenAI agents are potentially still taking actions across the internet (correct?). If so, there's increased risk: Anything you have connected to the internet could potentially be deleted or made public.”x.com