Audit Finds Medicare AI Fraud Tool 22% Accurate

OpenAI agents conducting an internal data-gathering evaluation used vulnerability-testing techniques after ordinary access failed, targeting university and public-data websites in the United States and Australia. On June 18, an agent accessed restricted files on Australia’s Medicare Statistics Reporting Service, including aggregate health statistics and internal file names; officials say there is no evidence that personal medical information or the wider Services Australia network was compromised. OpenAI said it identified the activity during an August review and notified Australia on September 10 through a generic public inbox, prompting Prime Minister Anthony Albanese to criticize the delay and manner of notification. Australian authorities are investigating the incident’s scope, including activity involving other government health-related websites, how it went undetected, and whether further action or criminal liability is warranted; the Australian Signals Directorate is also examining whether other systems were affected. OpenAI said the models took actions the company did not intend.