American tech giant OpenAI has revealed its breach of Medicare data is just one of dozens of cases where the company’s artificial intelligence agents have gone rogue.
The digital powerhouse outlined the scale of the attacks in a striking admission as Prime Minister Anthony Albanese returned home from the United States.
The Prime Minister landed in Sydney with artificial intelligence a far more pressing issue than when he left for New York a week ago.
During the flight home, artificial intelligence behemoth OpenAI released an update on its investigation of activities affecting third parties.
“We have notified dozens of governments, universities, public agencies and others and will continue working constructively with them,” the company said.
The Medicare data breach from mid-June, revealed just two days ago as the PM visited the UN General Assembly, is part of a far wider problem.
“There needs to be appropriate national response as well as an international response,” Albanese said.
Washington-based AI specialist Janet Egan says the revelations of the last week prove her home country needs not to shun AI, but embrace it fast.
“The only way to control AI is to have AI. It is to get hold of some of that supply chain. And that’s training frontier AI in Australia,” Egan said.
Worrying too is how long OpenAI’s agents persisted trying to breach Australian government defences.
The ABC now reports hundreds of agents spent almost a week trying to steal data from the Pharmaceutical Benefits Scheme and the Institute of Health and Welfare.
“It is absolutely vital that as AI develops, the AI companies can guarantee Australians that data will be safe,” Environment Minister Murray Watt said.