OpenAI will front up to a parliamentary grilling after one of its artificial intelligence agents went rogue and hacked a Medicare website.
A spokesperson for the tech giant on Monday confirmed its chief strategy officer Jason Kwon will fly from the US to Sydney to appear before a committee of MPs and senators on October 6.
The news came after OpenAI chief executive Sam Altman and Anthropic boss Dario Amodei were summoned to face a separate Senate inquiry into data centres holding hearings on Thursday.
Mr Altman will not heed that summons, OpenAI confirmed on Monday, due to the difficulty of wrangling executives to travel at such short notice, but the company will keep in touch with the Senate committee on possibly attending future grillings.
Anthropic will also miss Thursday's hearing but likewise send an envoy to Sydney on October 6, the Sydney Morning Herald reported on Monday.
AAP has contacted Anthropic for comment.
OpenAI has been slammed in recent days for keeping Canberra in the dark for months about one of its models going rogue and breaching a Medicare statistics portal.
It was the first known instance of an AI agent acting without its parent company's permission to break into an Australian government data system.
Experts had been working for months to crack down on potential AI break-ins before one happened, Public Service Minister Katy Gallagher said on Monday.
They included investing staff in AI work within the communication and industry departments, along with a $160 million cybersecurity package for Services Australia in the last budget.
"It's not like we waited for this event before we put a lot of time and effort and investment into safety in the AI world," Senator Gallagher told reporters.
Prime Minister Anthony Albanese on Monday stressed AI was extremely useful despite the security challenges it posed all over the world.
"It can boost productivity, can boost economic growth, it can give us health breakthroughs, but we need to make sure that human beings are in charge," he said.
A warp-speed government review is examining how to legally punish OpenAI for the breach and how Canberra should overhaul its relationship with the technology.
The company's executive poised to front the hearings, Mr Kwon, has posted on social media about the urgency of governments fortifying their cyber defences.
"This is work we can begin now - and it cannot wait," he wrote, before the Medicare breach was revealed.
In a submission to the joint parliamentary committee, made after the company had disclosed the breach but before the public knew, OpenAI encouraged the government to keep partnering with them on cybersecurity.
It advocated for Canberra to use AI as a defensive tool while maintaining human oversight and shared input on consequential decisions.
