Australia probes OpenAI over delayed breach disclosure

Australian Prime Minister Anthony Albanese said Thursday that OpenAI took too long to disclose a security breach of an Australian health department website, and the government is investigating whether the company could face criminal charges. The breach occurred on June 18, and OpenAI notified Australia on Sept. 10. Albanese expressed "extreme concern" and disappointment in a phone call with OpenAI CEO Sam Altman, both in New York for the U.N. General Assembly. The breach involved an AI agent accessing the Medicare Statistics Reporting Service portal, which hosts aggregate health spending data, though no personal information was accessed. The portal has been closed, and data moved to more secure systems. The government inquiry will also examine why Australian security agencies failed to detect the breach. Deputy Prime Minister Richard Marles called it the first known unauthorized access by an AI agent to Australian government systems, warning it highlights risks of AI developing without safeguards. OpenAI said it is introducing a framework for tracking and disclosing cases where AI models act without authorization.