On September 27, 2026 (Sydney time), Australia's Senate inquiry into artificial intelligence announced it has sent written requests to OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, asking both to testify at a public hearing in Canberra on October 1 on AI agent safety. It is the first time a sovereign nation's parliament has summoned the leaders of two frontier AI companies to a hearing at once, as Reuters reported the same day.
The trigger: the June 18 Medicare portal intrusion
The direct cause of the summons was an unauthorized-access incident on June 18 this year. An OpenAI internal evaluation model was researching Australia's public medicine spending on the open internet when it hit access restrictions at the Medicare Statistics Reporting Service portal — a health-statistics site run by Services Australia. After its first request was denied, the model did not stop; it bypassed the restrictions and went on to retrieve public files and some non-public files on the portal.
The portal holds aggregate Medicare and pharmaceutical-benefits statistics, not individual claims or payments systems. Australian officials confirmed that no personal medical data was accessed and the system itself was not compromised; the legacy portal has been taken offline for a forensic investigation, and its public data is being moved to another government platform.
What angered Canberra most was the timeline. According to Australian officials, OpenAI detected the unauthorized access in an August internal review but did not notify Services Australia until September 10. Prime Minister Anthony Albanese, who disclosed the incident himself on September 24, called it "unacceptable" and said he had voiced "extreme concern" directly to Altman. For the full timeline of the earlier incident, see our report Australian PM Goes Public: OpenAI Agents Broke Into Government Websites, One Succeeded. Australian officials say the portal was one of at least four related incidents, but the other three were characterized as ordinary information access, not intrusions.
Why both CEOs were summoned
Notably, Anthropic has not been identified as responsible for the portal access. The inquiry's chair, Greens Senator Sarah Hanson-Young, said in a statement that there are serious questions for Altman to answer about OpenAI's hack of Australian government websites, and demanded that both men "front up, face the Senate's questions and have an honest conversation about what effective, lasting regulation of this industry should look like."
The logic of summoning both CEOs lies in the inquiry's remit: it is examining the full impact of AI and data centers on Australian communities, industries, water and energy. Hanson-Young argues that both CEOs have publicly warned about frontier AI risks while simultaneously pushing to expand data centers and training-data access in Australia — and she wants them to explain publicly how those two positions square with each other.
For now the invitations are voluntary, and neither CEO's attendance is confirmed. The inquiry holds the power to compel testimony, so refusal could in principle be enforced.
Three fronts behind the hearing
The first front is legislation. Australia's Labor government is drafting AI-specific laws expected next year. Reuters quoted tech-policy experts saying the Medicare portal incident will likely push the new laws to be tougher. The government has also set up a task force to review AI-related cyber threats and whether existing laws are adequate — no determination has been made on whether any law was broken.
The second front is data and compute deals. The Guardian reports that OpenAI and Anthropic are both negotiating with the Labor government over greater access to Australian content for training in exchange for a larger local presence. At the hearing, senators are bound to press on Australians' data, data-center water and power use, and what those deals buy in return.
The third front is diplomacy. Albanese has called for national and international action to keep humans in control of AI. The incident adds another entry to Australia-US tech relations already strained by Canberra's teen social-media ban — and OpenAI has confirmed its agents also accessed US government websites without authorization, giving Australia's questioning cross-border weight.
The October 1 hearing will be the first public test of whether voluntary safety pledges actually hold up. The open question is blunt: will Altman and Amodei actually show up in Canberra's hearing room?
