Australia plans to introduce mandatory audits for AI model developers

The Australian government, led by Prime Minister Anthony Albanese, has introduced a draft law that establishes mandatory safety rules for companies developing frontier AI models. The initiative aims to create a legal framework for overseeing the most powerful artificial intelligence systems.

According to the proposed document, a special regulator will be established in the country with the authority to mandate independent model audits. Companies whose systems fall under the frontier category will be required to provide access to their developments for third-party testing. In the event of a breach of established safety standards, the bill provides for the imposition of strict penalties.

Canberra authorities emphasize that this measure is a priority for the domestic market. The government seeks to solidify national safety standards before actively participating in international negotiations on AI regulation. This will allow Australia to act from a position of established legal policy when discussing global rules of the game for technology developers.

For practitioners and companies working in the AI sector, this means a potential change in the conditions for accessing the Australian market. The introduction of mandatory audits may require developers to revise their model release preparation processes to comply with the regulator's transparency and safety requirements. Further details on the implementation of the law will depend on the final text of the document and the operational mechanisms of the future supervisory body.