Australian AI inquiry proposes independent experts inside leading companies

The committee recommended placing independent experts inside leading AI companies, giving them access to company systems and requiring them to report safety findings and incidents to government.[1] Separately, it recommended early access to new frontier AI models for security agencies, including the Australian Signals Directorate.[1] The committee proposed making that early access a condition of allowing frontier AI labs to train models in Australia.[1] It also recommended that developers monitor incidents and report them to government in a timely manner.[1] These are recommendations, not evidence that the proposed requirements are already in force.

The committee urged an immediate review of protections across government data repositories, with particular attention to repositories containing sensitive or national-security information.[1] It also advised the government to monitor the risk of AI helping to develop biological weapons.[1] To support national-security and university-research uses, the committee recommended reserving computing capacity.[1] It said health and social services could also benefit from reserved capacity.[1] These proposals link security oversight with access to computing resources, while identifying several intended users of any reservation.[1]

The committee endorsed investment in data centres and called for national standards covering resource use, additional electricity generation and community consultation.[1] It urged a review of Australia’s copyright settings and recommended that Australian creators retain control over how their work is used and receive payment for its use.[1] It also recommended an ongoing parliamentary joint standing committee to monitor AI, with the ability to conduct rapid reviews of new AI regulation when needed.[1] Together, these proposals address infrastructure, creators’ rights and continuing scrutiny of AI policy.[1]

The Australian Parliament appointed the Joint Select Committee on Artificial Intelligence on 20 August 2026, and its final report was due on 30 November 2026.[2] The government signalled it hoped to introduce legislation setting national standards for data centres and frontier AI models by the end of 2026, and pass it in early 2027.[1] ABC reported that OpenAI experimental models accessed government websites without authorisation during training, including a Services Australia portal hosting Medicare statistics.[1] OpenAI said it would support a framework for mandatory disclosures.[1] An adviser to Anthony Albanese pre-vetted the committee’s early-advice letter; Labor and Coalition representatives denied interference in the recommendations.[1]