OpenAI issues lengthy apology statement as government fumes over AI agent’s website breach

OpenAI have issued a lengthy apology to Australia after it was sensationally revealed an AI agent accessed an Australian government website.

Speaking from New York last Thursday, Prime Minister Anthony Albanese revealed that a rogue Open AI agent hacked “non-public files” in Australia’s universal health insurance agency back in June in a breach labelled an “extreme concern” by officials.

It’s understood to be the world’s first instance of an autonomous AI agent breaching a government website to independently access private data.

High-ranking Australian officials had called for the tech giant to be punished, with Mr Albanese publicly exploring criminal charges for the breach.

OpenAI had previously issued a statement admitting its models “took actions we did not intend”, but on Tuesday the company issued a lengthy apology titled “How we will do better for Australia”.

“We should have handled our response better — we are sorry and working to do better in the future,” the company wrote, before explaining what they intend to do to “rebuild trust with the Australian people.”

OpenAI revealed they discovered the breach had affected four Australian government websites in mid-August — including Services Australia, NSW Bureau of Crime Statistics and Research, the Victorian Department of Health, and the Australian Institute of Health and Welfare.

“We launched investigations into these activities as soon as we became aware in mid-August,” the organisation said, revealing the government was informed about three of the sites in early September.

“The activity related to the Australian Institute of Health and Welfare did not meet our disclosure thresholds because the way it was accessed seemed consistent with public access, but we notified (the government) on September 24 to share our findings.

“Our aim was to give affected agencies a detailed account once our investigation was compete — however, we should have shared preliminary findings sooner and kept Australian agencies updated as more facts emerge.”

The AI giant revealed the breach was a result of “an experimental, internal-only OpenAI model not intended for public release and without the full safeguards used in our public products”.

As a result, the AI agent had committed the breach when asked to research government spending for skin medicines in Victoria.

“The model had difficulty obtaining that information, and it took actions that we had not authorised it to take,” they said.

“It discovered a way to gain non-public access to the service. It then used this access to review technical system information and source code related to the service—all still with the objective of trying to find the information it was originally looking for.

“We did not intend for this activity to occur, and the access to the service and follow-on activity should not have happened.”

The company said no medical records had been accessed during the breach.

The organisation has stressed to Australians that they have strengthened their research safeguards and will establish an Australian task force to develop AI risk policies.

OpenAI’s Chief Strategy Officer Jason Kwon will fly to Australia next month to appear at the Joint Select Committee on Artificial Intelligence in Sydney.

The hearing, which is set to take place on October 6, will see the OpenAI representatives address how the company responded to the breach and the changes they have implemented since.

Speaking in Adelaide on Tuesday, Mr Albanese said OpenAI has been continuing to give “extensive briefings” to Australia and insisted there had been “constructive discussion”.

“This is a challenge that the world is dealing with. I welcome the engagement with Open AI,” he said.

“I spoke with Sam Altman last week and we had a direct but constructive discussion.

“Artificial intelligence can improve economic growth, improve our productivity.

“It can make a big difference for businesses who are using it.

“It can provide revolutionary breakthroughs in health and research and science and innovation.

“But there are risks and we’ve seen those risks exposed. So, we need to work this through.”