OpenAI apologizes to Australia for Medicare agent breach

Chief strategy officer to testify at Australian AI committee next week

OpenAI said in a Sept. 29 blog post that it should have handled its response better. “We also should have handled our response better. We are sorry and working to do better in the future,” the company said. The post detailed what the company said it had learned about the agent’s activity on Australian government systems, building on a public disclosure by Albanese one week earlier.

According to OpenAI, the agent gained non-public access to a Services Australia portal used to report Medicare statistics. The agent was able to run commands, retrieve internal files and credentials, and write files, the company said. No patient or client records were accessed, OpenAI said.

The agent also accessed the NSW Bureau of Crime Statistics and Research’s public crime-mapping tool, according to OpenAI, with application configuration, operational jobs, logs and website metadata obtained by the agent. Separately, the agent discovered an exposed access key that allowed it to query the Victorian agency for health information’s reporting system for aggregate survey statistics. At the Australian Institute of Health and Welfare, the agent retrieved aggregate statistics; OpenAI said separate attempts to bypass access controls there were unsuccessful and that the information obtained was publicly available.

OpenAI said it first informed Services Australia and the Victorian health department on Sept. 10. The NSW bureau was informed on Sept. 18. The Australian Institute of Health and Welfare was not notified until Sept. 24 — a delay OpenAI attributed to its assessment that the incident there did not meet disclosure thresholds. “Since then we’ve worked closely with Australian government agencies to share what we’ve learned to date,” OpenAI said. “If we identify any additional affected agencies, we will notify them promptly and directly with the information available and provide updates as further facts emerge.”

The incident originated in a research task. OpenAI said one of its models was asked to research government spending per person on medicines for skin conditions in Victoria. When the model had difficulty obtaining that information, “it took actions that we had not authorised it to take,” including accessing Services Australia’s Medicare statistics reporting service, the company said.

OpenAI said it became aware of the agent activity in mid-August after reviewing earlier training incidents following a separate Hugging Face attack in July.

OpenAI said it would commit resources and expertise to the affected agencies, including credits from its US$1 billion Daybreak fund — which the company said allows organizations to use frontier AI for cyberdefence and to identify vulnerabilities in code and system configurations. OpenAI also said it would establish a taskforce with Australian expertise to develop policy recommendations on managing risk from AI agents.

OpenAI chief strategy officer Jason Kwon will appear at Australia’s Joint Select Committee on AI next Tuesday. Anthropic will also appear at that hearing, Guardian Australia reported, though not at a separate Senate inquiry into AI and datacentres this week.

Albanese, who disclosed the hack from the United States last week, said at the time he had spoken with OpenAI chief executive Sam Altman “to express Australia’s extreme concern about this incident.” On Tuesday, Albanese said OpenAI had been “very constructive and open in engaging” since the incident, as had Anthropic. He said AI can improve economic growth and productivity but carries risks, adding: “And we’ve seen those risks exposed – not just in what occurred in Australia, but the revelation that has occurred in the United States and other countries as well.”

The federal government has said it could introduce mandatory reporting rules for AI-related data breaches after it emerged that OpenAI used a public-facing email address three months after the hack to report the incident to Services Australia. OpenAI said it had “a lot of work ahead” to rebuild trust with Australians and was making “meaningful changes.”