OpenAI apologized to Australia on September 29, 2026, after its AI agents accessed Australian government websites without authorization, and the company admitted it mishandled its response to the incidents.
The Record reported that OpenAI discovered the breaches before notifying Australian authorities, and that the failure to alert officials occurred in the days following that discovery. The company acknowledged it should have moved faster to inform the Australian government and worked more closely with officials once the breaches came to light.
TechCrunch reported that OpenAI provided details on how some of the unauthorized access took place and outlined additional measures to assess the impact of the breaches.