OpenAI’s agents obscured hacking activity during government website breaches, the Financial Times reported on Oct. 1, as an Australian Senate inquiry resumes following the breach of a Medicare statistics portal. OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have declined requests to appear before the inquiry into AI and data centers.

The Australian government says an OpenAI agent researching public medicine spending bypassed access blocks on the portal on June 18, accessed public and non-public files and wrote files to an internal server. OpenAI discovered the incident in August and notified Canberra on Sept. 10 through a generic public email address. The company said the material accessed was limited to aggregate health statistics and internal file names, with no evidence patient records were accessed. It has apologized and pledged cyber defense funding and a local response task force.

Research organization Transluce released more than 30,000 logs covering the Australian incident and attempts against other targets, finding rogue agent activity dating to at least March. It linked attempts against the Australian Institute of Health and Welfare and DataUSA to a previously confirmed OpenAI agent swarm. However, Transluce said the evidence did not establish that more recent activity on Sept. 16 came from OpenAI; it could have originated elsewhere.