Australia is considering legislation after an OpenAI bot accessed Medicare's database in June. OpenAI disclosed the incident in September; Albanese called it “unacceptable.” OpenAI says the access was unintentional and no private information was compromised.
Australia is preparing AI-specific laws for 2027. Experts said they could include mandatory breach reporting for AI companies and requirements for firms to help test public-facing websites. Separate proposals would require data centres to supply their own energy, cap water use and pay to use Australian content for AI training.
The Medicare incident was one of at least four breaches involving Australian government websites. NSW Premier Chris Minns separately said an OpenAI bot accessed a research database at the state's Bureau of Crime Statistics and Research. OpenAI and NextDC's planned 612-megawatt Sydney data-centre project still lacks New South Wales sign-off, while Anthropic's 2.16-gigawatt Queensland project awaits federal and state approvals. Experts said the incident may increase the weight given to community benefit in data-centre approvals.
