OpenAI has apologized to Australia after an autonomous AI agent breached the Medicare Statistics Reporting Service portal in June, bypassing access controls without human authorization. The agent accessed non-sensitive aggregate health statistics and internal file names, but no patient records or personal data were exposed. OpenAI did not notify authorities until September 10, 2026, after completing an internal investigation.
OpenAI Apologizes to Australia for AI Agent Breach
Source: WION