Australia needs AI frontier model access deal with US: Ted O’Brien ‘Prisoner’s dilemma’: AI companies accused of putting ‘capability ahead of safety’The Coalition says Australia needs access to cutting-edge AI models to “protect” itself against foes as well as the technology itself.Joseph Olbrycht-Palmer and Nathan Schmidt4 min readOctober 8, 2026 - 11:29AMNewsWireAI companies are putting “capability ahead of safety” as the Albanese government seeks to regulate frontier models, a leading minister has warned. Assistant Science Minister Andrew Charlton made the case for government regulation of frontier AI models, over market-led alternatives, during a speech before the Sydney Trust and Safety Festival on Thursday.It comes days after OpenAI chief strategy officer Jason Kwon fronted a joint parliamentary inquiry following the breach by one of the US company's models of a Medicare data portal.Globally, leaders have disagreed over how cutting-edge models should be regulate, with US President Donald Trump disagreeing with Silicon Valley over the need to slow the pace of new development.On Thursday, Mr Charlton said new technology should be built on trust, but that there was “every indication is frontier labs are putting capability ahead of safety”.Assistant Science Minister Andrew Charlton made the case for government regulation of frontier AI models. Picture: NewsWire / Martin Ollman.“Just listen to the people who know these companies best … last month, Jacob Coxon, who had worked at both OpenAI and Anthropic, quit the industry altogether. Others you may likenewsnews“His verdict was blunt: ‘Neither company is acting responsibly’.”Mr Charlton said voluntary regulation and codes were “fast and flexible, but they fail the incentive test” when it came to frontier AI. “You cannot ask firms to volunteer against their own competitive interest amid a manic race involving trillions of dollars and then be surprised when they do not,” he said. “Detailed prescriptive rules can offer clarity, but at the pace this technology moves, which is months not years, those rules could be out of date before the ink is dry.“Systems regulation places the onus on companies to build and run a rigorous process for finding, testing, reporting on and managing the risks of their systems, and then holding them accountable for whether that process works.”Mr Charlton said a such a system would not only hold companies to account when something went wrong, but also “detect risk and prevent incidents occurring”.OpenAI chief strategy officer Jason Kwon fronted a joint parliamentary inquiry on Tuesday. Picture: NewsWire / Nikki ShortHowever, he admitted regulations would not stop “hostile states or criminals”, and would be supplemented by “technology Australia has a role developing at home”. That includes a focus on developing AI talent and training onshore.The government is expected to announces its first-ever national standards for AI shortly.‘Two-track’ raceMr Charlton said the leading companies were in a two-track race: one with each other, and the other with China.“And for now – everyone has calculated that the biggest reward comes from being first, even when it’s not clear what the higher purpose of coming first is.”Mr Charlton drew on the economics theory of the “prisoner’s dilemma”.“Every lab would be better off if they slowed down together,” he said.“But so long as profits are internalised, and costs externalised, any lab that slows down alone simply hands the lead to one that does not.“In fact, each company tells itself a story in which the frontier is safer in its own hands.“ … And that is why the government needs to step in.”Mr O’Brien says Prime Minister Anthony Albanese should have locked in a deal while in the US last month. Picture: NewsWire / Handout / Prime Minister’s OfficeMr Charlton said lessons from the Medicare incident would shape AI standards legislation. “A race to the bottom is not what competition looks like when it works. It is what competition looks like when it fails,” he said.“And market failure is not a reason to give up on markets. It is the reason policy exists.”In Australia, ministers from across multiple portfolios were “taking responsibility for AI harms”, including through the Digital Duty of Care, Mr Charlton said. Mr Charlton said AI posed a distinct regulatory challenge owing to the pace of change, information asymmetry, and that it is hard to recall. Risks from AI could also emerge later and in “unexpected and unintended ways”, Mr Charlton said.“Traditional software is built from rules people write and can inspect. Frontier AI is not.“ … its abilities emerge in ways even its creators cannot predict.”‘Do a deal’It comes as the opposition is calling for a deal with the US giving Australia “preferential access” to cutting-edge artificial intelligence models.In a speech on Wednesday night canvassing the fierce AI race between the US and China, opposition foreign affairs spokesman Ted O’Brien said Australia needed access to American frontier models to protect itself.Opposition foreign affairs spokesman Ted O’Brien is calling for a deal with the US to access American frontier models. Picture: NewsWire / Martin Ollman“The recent incident involving AI agents independently hacking into Medicare data is a potent reminder of such risks,” he told the Australia-China Relations Institute.“AI isn’t a choice – it’s a reality. It is already here and expanding rapidly.”He said the “world needs safeguards” to not only “protect against AI being misused by state and non-state actors but to mitigate the horrifying prospect of AI being weaponised by itself”. “The answer – we must do a deal with the United States,” Mr O’Brien said.“Such a deal could start with a government-to-government framework that affords Australia early and preferential access to AI frontier models and a guarantee of continued access.More CoverageOpenAI boss ‘unaware’ of Medicare breachWilliam Elliott, Joseph Olbrycht-Palmer, Nathan Schmidt‘Our time’: Albo’s plan to cash in on AINathan Schmidt and Jack Nivison“Australia’s immediate priority is to fortify our digital defences and shield the critical infrastructure underpinning modern society from AI-driven cyber threats.“We can use AI as a shield to counter AI being used as a sword.”He also said that is what Prime Minister Anthony Albanese should have sought while in the US for the UN General Assembly in September.Join the conversationAdd your comment to this storyTo join the conversation, please log in. Don't have an account? RegisterJoin the conversation, you are commenting as LogoutMore related storiesEnvironmentShocking stat about Australian wildlifeSome of Australia’s rarest animals are vanishing at an alarming rate, with conservationists sounding the alarm as the threat from bird flu grows.Read moreTravel‘Spectacular’ first for Aussie travellersVisitors to one Australian state could come face-to-face with a rare natural phenomenon in national parks for the first time.Read moreAnimalsWorld’s oldest orangutan dies at Aussie zooOne of the world’s oldest Sumatran orangutans, once visited by iconic conservationist Jane Goodall, has died at an Aussie zoo aged 56.Read more