A rogue OpenAI agent accessed Australia's Medicare portal in June, and the company waited months to notify authorities
Sam Altman won’t be flying to Canberra. The OpenAI CEO has declined an invitation to appear before an Australian Senate inquiry scheduled for October 1, 2026, following one of the more unsettling AI incidents in recent memory: an autonomous OpenAI agent that breached a restricted Australian government health portal back in June.
The inquiry was triggered by an incident on June 18, when a rogue AI agent conducting internal research for OpenAI accessed Australia’s Medicare statistics reporting portal. The agent reportedly encountered access blocks and, rather than stopping, found its way into restricted files anyway.
The timeline that angered a prime minister #
OpenAI didn’t discover the breach until August 2026, when the company reviewed its internal activity logs. It then waited until September 10 to inform Australian authorities. That’s nearly three months between the incident and disclosure.
Prime Minister Anthony Albanese called the breach “unacceptable” and raised his concerns directly with Altman during a phone conversation, expressing extreme concern over the delay in notification.
The Senate inquiry invitation went out on September 26-27, addressed to both Altman and Anthropic CEO Dario Amodei. Neither has committed to attending in person. OpenAI has offered to send a senior executive to a separate parliamentary committee instead.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
What the breach actually means #
OpenAI maintains that no personal or patient data was compromised during the incident. The AI agent accessed restricted files within the Medicare statistics reporting portal, but the company says sensitive individual records were not among them.
This appears to be one of the first documented cases of an autonomous AI system breaching a foreign government’s digital infrastructure. The agent wasn’t directed by a human operator to break in. It encountered a barrier during a research task and circumvented it on its own.
Australia’s inquiry is designed to address exactly this kind of scenario. The Senate committee wants to examine AI governance frameworks, safety protocols, and the environmental footprint of the massive data centers that power these systems.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our