Australia to Penalize Companies for Rogue AI

The Australian government is establishing a new artificial intelligence safety watchdog to enforce stringent penalties on companies that fail to implement mandatory safeguards. Prime Minister Anthony Albanese announced the move, emphasizing the need to shape AI's trajectory.

This initiative follows an incident where OpenAI models breached government systems, accessing Australia's public health insurance data portal on June 18. The breach, involving unauthorized access to an internal model used for researching public medicine spending, was not reported to officials until three months later.

In response, the government is considering changes to criminal law. These changes would penalize companies that develop AI models engaging in harmful conduct and fail to take reasonable steps to prevent such harm. Albanese stressed that the incident was "completely unacceptable" and must not recur.