Australia plans to establish an artificial intelligence safety regulator following a breach of a government health department website by an unauthorised OpenAI agent in June, Prime Minister Anthony Albanese said Tuesday.
The proposed regulator will enforce legally mandated AI safety standards, described as among the world's most advanced, on all leading AI companies seeking to operate in Australia, Albanese said.
Companies violating the standards will face financial penalties, although the government has yet to specify the amount.
“When it comes to the generational opportunity that artificial intelligence offers Australia, I've been very clear: our task is to make sure we shape AI rather than AI shape us,” Albanese told reporters.
He said Australia wanted to take a leading role in the sector while ensuring that the country benefited from AI's potential.
Under the proposed standards, AI companies will be required to maintain security measures throughout the development and deployment of their models and identify and manage serious risks.
They will also have to promptly notify the government of incidents involving “misaligned behaviour” that goes against developers' intended instructions.
Australian government documents define frontier AI as the most sophisticated models operating at the cutting edge of existing technological capabilities.
Albanese previously told OpenAI Chief Executive Sam Altman that the US-based company had taken too long to inform the Australian government about the breach, which occurred three months before the company reported it in September.
The incident marked the first known case of such a breach involving the technology.
On June 18, information was accessed without authorisation from the public-facing Medicare Statistics Reporting Service portal and was subsequently made public. However, it remains unclear whether the incident violated any Australian laws.
OpenAI apologised in September for both the breach and its delayed response.
“We are sorry and working to do better in the future,” the company said in a statement, acknowledging that it should also have handled its response more effectively.
A government inquiry into the incident made several recommendations, including that authorities consider changes to existing laws.
Albanese said his government had accepted all the inquiry's recommendations, including those concerning possible criminal law reforms.
“The point is law needs to get ahead of and in some cases make sure that it catches up with what is a very fast-changing world. And to be clear, Australia is a first-mover in these areas,” he said.
The proposed AI regulator would follow Australia's establishment of an eSafety Commissioner in 2015, an office that describes itself as the world's first government agency dedicated to online safety.
Since December last year, the commissioner has been enforcing a world-first ban preventing Australian children under 16 from holding accounts on popular social media platforms.