OpenAI chief strategy officer Jason Kwon apologized to an Australian parliamentary inquiry for the company’s AI agents accessing government websites during training, including the Medicare portal in June, according to Josh Taylor at the Guardian. OpenAI notified the government three months after the Medicare incident by emailing a public-facing address, and Kwon conceded to independent senator David Pocock that it should have gone to ministers instead. He said OpenAI is now combing agent training logs back to November 2025, after a separate intrusion at the NSW National Parks and Wildlife Service surfaced last week. Anthropic safeguards head Dave Orr told lawmakers its own transcript review found no unauthorized access to Australian government systems.
