Australia should require AI companies to submit to independent safety checks, reserve computing power for local use and give security agencies early access to new models, a parliamentary committee has recommended.
The bipartisan inquiry examining artificial intelligence has presented "early advice" to Prime Minister Anthony Albanese in the wake of the OpenAI breach, in which autonomous agents gained unauthorised access to Medicare statistics.
The committee has endorsed investment in data centres, backed by national standards covering resource use, additional electricity generation and community consultation.
It has also urged a review of Australia's copyright settings, which would be at odds with creative groups seeking stronger enforcement of existing protections rather than changes to accommodate the AI industry.
Prime Minister Anthony Albanese will make the decision about whether to implement many of the recommendations. (ABC News: Matt Roberts)
The recommendations were outlined in a letter sent to Mr Albanese on Monday night, signed by the committee's Labor chair Jo Briskey and Liberal deputy chair Jonno Duniam, with a final report due at the end of November.
The federal government has signalled it hopes to introduce legislation outlining national standards for data centres and frontier AI models by the end of this year and pass the laws through parliament in early 2027.
Push for independent experts inside AI companies
The committee has recommended the government put independent experts inside leading AI companies that would have access to their systems and a duty to report safety findings and incidents to government.
Security agencies like the Australian Signals Directorate should receive early access to new models as a condition of arrangements allowing frontier AI labs to train them in Australia.
The government has been urged to immediately review protections used across its data repositories, particularly those holding sensitive and national security information.
It has also been advised to monitor the risk of AI helping to develop biological weapons.
Alongside those safeguards, the committee has questioned whether existing copyright rules were delivering for Australian creators or supporting local AI development.
Copyright rules under scrutiny
The committee noted current copyright arrangements were "not facilitating" licensing deals for Australian culture and content.
"The status quo sees the work of Australian artists, musicians and media outlets used overseas without adequate control or payment," the letter said.
They also acknowledged frontier AI labs had indicated existing copyright settings were a "barrier" to undertaking training in Australia.
The committee recommended the government "examine" current copyright settings and prioritise ensuring that Australian creators "retain control over how their works are used, that they are paid, and that AI systems reflect Australian values and culture".
But Australian Recording Industry Association (ARIA) chief executive Annabelle Herd told the inquiry last week that existing licensing systems could support AI training without rewriting copyright protections to suit foreign companies.
"They should come to the negotiating table, not to the prime minister's office," she said.
OpenAI breach drives calls for oversight
Last month it was revealed OpenAI's experimental models accessed government websites without authorisation during training, including a Services Australia portal hosting Medicare statistics.
OpenAI has apologised for both the incident and its handling of the response, which included a months-long delay in detecting the issue and notifying the government about it.
The letter from the committee said the OpenAI breach demonstrated the need for "appropriate and robust safeguards" to protect Australian data and those who own and use it.
It was recommended the safeguards apply to frontier AI developers deploying models and agents in Australia, ensuring there was "proper and transparent accountability and oversight mechanisms" along with "well-established practices to deal with incidents that may occur in the future".
Frontier labs should also be required to monitor and report incidents to the government in a "timely manner".
"This should include an obligation or duty to monitor risks and clear reporting deadlines," the letter said.
OpenAI chief strategy officer Jason Kwon last week said the company would "support a framework on mandatory disclosures".
Australia cannot simply be 'consumer' of foreign AI
It was recommended the government establish an ongoing parliamentary joint standing committee to monitor AI, which could conduct "rapid reviews" of new regulation when required.
This committee would also be reported to at "regular intervals" on serious events and incidents, including cyber incidents involving AI, unintended actions of agents and models as well as "misuse" of the technology.
The letter recommended reserving a share of computing capacity for national security and university research, with health and social services also among the potential beneficiaries.
"The overarching objective should be to ensure that Australia is not merely a consumer of AI technologies developed elsewhere, but a trusted, competitive and strategically capable participant in the global AI industry," the letter said.