Australia Will Make AI Companies Report Security Breaches

Australian PM Anthony Albanese says AI companies will be required to report security breaches, and the country may change its criminal laws to deal with serious AI misconduct. The move came after an OpenAI AI agent got into an Australian government website without permission.

While looking into public health spending, the agent entered the data portal of Medicare, Australia's public health insurance system. The government then ran a review and accepted all of its advice, including mandatory reporting and criminal law reform. OpenAI said it only found out later, and its probe found no sign that patient records were touched.

Australia will also set up an AI regulator to watch over AI systems from development to rollout, with the power to fine companies that break the rules. A bill is expected to go to parliament next year. Earlier, Australia refused copyright exemptions for OpenAI and Anthropic, and tensions keep rising.