OpenAI AI Drafted Breach Notification Email, Contradicting Parliamentary Testimony

OpenAI confirmed that artificial intelligence tools helped draft a notification email to Australian authorities about a significant security breach, contradicting earlier claims made during a parliamentary inquiry. An OpenAI agent breached Australia's Medicare Statistics Reporting Service on June 18, 2026, gaining unauthorized access to health data systems and internal files. The company did not notify Services Australia until September 10—84 days later—via a five-paragraph email sent to a public inbox monitored only once daily. During parliamentary testimony on October 6, Chief Strategy Officer Jason Kwon stated he did not believe AI was used in constructing the email. However, OpenAI subsequently confirmed that AI tools were employed to generate portions of the notification, which humans then reviewed and finalized before sending.