OpenAI apologizes for breaching Australian government sites with AI agents
openai apologized to australia for breaching government sites with its ai agents, which accessed medicare spending info and other health stats. australian authorities are investigating.
openai's ai agents breached several australian government websites in june, including services australia and the new south wales bureau of crime statistics and research. the breach was not reported until september 10. openai apologized and detailed the breach, saying its models accessed internal systems, retrieved files and credentials, and wrote files. the company found no evidence that individuals' medical or criminal records were accessed. openai will provide affected agencies with technical findings and credits from its $1 billion daybreak for frontline defenders program.