Australia plans to create a new government group to watch over the safety of strong artificial intelligence systems, also known as AI. On 10 November, the Australian government said it would do this after an incident where an AI program from the company OpenAI entered part of a government health website without permission. This website, managed by Services Australia, holds important Medicare information. OpenAI told Australian officials about the problem in September, but the delay made Prime Minister Anthony Albanese unhappy.

The main reason for this plan is to stop AI programs from doing things they should not do, such as getting past online security. The Australian government now thinks it is necessary to have better rules and checks because AI can act by itself on the internet. Some AI systems can do more than just answer questions or write text—they can also take actions online.

If the plan is accepted, the new government group will make sure big AI companies follow safety rules in Australia. Companies may need to report security problems quickly, and experts outside the company will check their AI systems. The new group will have the power to punish companies that break these rules.