Australia to Legislate Mandatory Safety Rules for Frontier AI Companies

Canberra, Australia — Australia plans to write binding safety and security duties for frontier artificial intelligence developers into its upcoming AI Standards legislation, with a new AI safety regulator empowered to levy significant penalties on companies that fail to comply, Prime Minister Anthony Albanese announced on Tuesday.

The duties would cover every frontier AI company that wants to operate in Australia, according to a joint media release issued by Albanese, Deputy Prime Minister and Minister for Defence Richard Marles, Minister for Home Affairs and Cyber Security Tony Burke and Minister for Industry and Innovation Tim Ayres. The government called the protections world leading and placed them within its AI in Australia's Interests agenda.

The release defines frontier AI as the most sophisticated models, those working at the outer edge of what today's technology can do. Such systems bring large opportunities, the government said, but also the greatest risks to the community when they are used unsafely, which is why it described safety and security as the base on which trust in them must be built.

Six requirements would be imposed on companies by the regulator. They would have to keep security controls in place at every point from building a model to deploying it, find and manage the serious risks their models carry, and open those models to independent evaluation. They would also need to report incidents to the Australian Government rapidly on a mandatory basis, keep a published safety and security framework to support good governance, risk management and transparency, and adjust to revised standards and rules as the technology moves on.

Enforcement would sit with a new, targeted AI safety regulator that the government will propose. That body would work with the existing Australian AI Safety Institute, which would lend technical expertise and oversight while international standards continue to develop. According to the release, the regulator would have the power to impose significant penalties on AI companies that breach the mandatory obligations.

The government said the package is consistent with recommendations made in the interim report of the Joint Select Committee on Artificial Intelligence. It expects the rules to flag risks from rogue AI agents sooner as they emerge, and to let authorities respond faster and in a coordinated way when a serious AI incident takes place.

Timing was deliberate, the release said. The government wants AI investors to come to Australia, but said protecting Australians comes first, so it chose to set out its safety approach before a series of major international talks on AI development and investment.

The ministers also framed AI as a source of large gains for the country, from lifting productivity and strengthening economic resilience to revolutionizing healthcare. Capturing those gains, they said, requires a social license for AI, keeping Australians safe and protecting the national interest. The release pointed to Australia's reputation as a world leader in digital safety standards and said its AI approach keeps Australian values at the forefront.

The obligations are not yet law. The government said the AI Standards legislation also deals with several other policy areas it has already outlined, and that fuller details of the bill will be released in due course, after it finishes its detailed work and consultation.

Anthony Albanese, Prime Minister of Australia, said: “Artificial intelligence presents enormous opportunities for Australia, but we must also be prepared for the new risks that come with rapidly advancing technology. These protections will ensure we keep Australians safe and secure our national interests, while seizing the generational opportunities that AI offers. Our world-leading AI in Australia’s Interests agenda aims to provide Australians confidence that frontier AI companies are taking steps to improve safety and security, and provide certainty for investors – boosting productivity, building economic resilience, while bolstering sovereignty.”

Richard Marles, Deputy Prime Minister and Minister for Defence, added: “Getting these settings right will help ensure Australia can harness the opportunities artificial intelligence presents for our economy, our national security and our sovereignty. The Albanese Government is committed to putting in place the protections that will allow Australia to be at the forefront of this technology to keep Australians safe.”

Tony Burke, Minister for Home Affairs and Cyber Security, said: “Those who want to cause us harm have access to the latest technology. We need to make sure we have the same, or better. But it’s not a free-for-all, there needs to be clear guardrails to protect Australians.”

Tim Ayres, Minister for Industry and Innovation, said: “Keeping Australians safe is at the heart of the Albanese Labor Government’s approach to artificial intelligence. The National AI Plan released last year put safety at the core of the Government’s AI agenda - capturing the opportunities of AI while keeping Australians safe. Legislating protections as part of the Australian AI Standards will build on the work of the National AI Plan, helping Australia spread the benefits of AI across the country, while making sure that AI systems are developed and deployed in Australia’s national interest. Strong protections will ensure that Australia has the capacity to shape our own technological future and build our economic resilience.”

The institute that would advise the new regulator is itself recent. Ayres announced the Australian AI Safety Institute on November 25, 2025, giving it the job of assessing emerging AI capabilities, sharing information and supporting timely action on potential risks, and it was later listed as a key action under the National AI Plan.

The rules would also land as AI infrastructure spending in the country climbs. In September, NVIDIA and Australian providers including AirTrunk, NEXTDC and CDC said they were targeting up to 2 gigawatts of AI computing capacity by 2027, and frontier developers serving that market could face the new evaluation and reporting duties.

The rapid incident-notification requirement follows a case that tested existing processes. Albanese said in September that an OpenAI agent had gained unauthorized access to a Medicare statistics portal run by Services Australia in June, but that OpenAI did not notify the government until September 10. He said a task force reviewing that incident would feed its findings into the government's AI standards work and refer them to a parliamentary committee examining artificial intelligence.