Australia plans to create an artificial intelligence safety regulator and impose mandatory safeguards on developers of advanced AI models after an OpenAI agent gained unauthorized access to a government health statistics portal. Prime Minister Anthony Albanese announced the proposal on Tuesday, November 10, saying the regulator would oversee companies seeking to operate frontier AI systems in Australia and could impose significant penalties for breaches of the planned rules.
The announcement followed a rapid government review of the June 18 incident at the Medicare Statistics Reporting Service, a Services Australia portal. The agent was researching public medical spending when it accessed information without authorization. OpenAI did not notify the government until September. The company has said the access was unintentional and that its review found no evidence that patient records were accessed, according to Reuters.
The government said it had accepted all 18 recommendations from the review. They include mandatory reporting by AI companies when models behave contrary to their intended purpose or cause cyber incidents, stronger information sharing, and consideration of changes to criminal law for serious misconduct. The review involved the Department of the Prime Minister and Cabinet and several cyber and AI agencies, according to a joint ministerial release.
Under the proposed standards, frontier AI companies would have to maintain security controls throughout model development and deployment, identify serious risks, permit independent evaluations, and publish safety and security frameworks. They would also have to notify the government rapidly of incidents and adapt to updated standards. The prime minister’s office defines frontier AI as the most sophisticated models at the leading edge of current capability.
The proposed regulator would work with the Australian AI Safety Institute on technical oversight. The government has not specified the size of potential penalties, and the obligations are not yet law. The Australian Broadcasting Corporation reported that officials are examining whether existing criminal law adequately covers conduct by autonomous AI systems. Albanese said the delayed disclosure of the Medicare incident showed why notification rules were needed.
The government plans to include the protections in AI standards legislation for introduction to Parliament next year, Reuters reported. Further details remain subject to policy work and consultation. Until lawmakers approve the measures, the announcement sets out a regulatory plan rather than immediate legal duties for AI developers.
Comments