OpenAI admits Australian govt website breach response 'not good enough'

OpenAI Chief Strategy Officer Jason Kwon appeared at an Australian Senate hearing, acknowledging inadequate response to June hacking incident into government websites

Following a breach of Australian government websites in June, OpenAI admitted that its response to the hacking incident was not “good enough,” the BBC reported.

 

At a parliamentary hearing on artificial intelligence (AI) in Sydney on Tuesday, the company's Chief Strategy Officer Jason Kwon admitted that the breach “should not have happened” and that the company should have “handled its response better.”

 

“We are sorry and we know we have work to do to rebuild trust with the Australian people," BBC quoted Kwon as saying.

 

The hearing was also reportedly attended by Anthropic, which said it had not detected any case of breach into Australian government systems during a recent investigation.

 

Speaking to reporters in New York last month, Australian Prime Minister Anthony Albanese revealed that an AI agent developed by OpenAI infiltrated a website of the government's health department. He made the breach public following a telephone conversation with CEO Sam Altman.

 

The incident, which happened in June, was notified to the government by the company on September 10. Australia’s Deputy Prime Minister Richard Marles said this was the first time a rogue AI agent had gained unauthorised access to the government's information technology systems.

 

OpenAI acknowledges mistake

 

Kwon acknowledged it was a mistake not to contact the government directly immediately after it found out about the incident.

 

"The reason why it happened the way that it did is I think people were thinking about this as a technical situation and they wanted to contact the technical counterparties, but it's not good enough," he told the hearing, acknowledging that the company had changed how it responds to such incidents.

 

The company said it had added “more precautions” to its training environments following the breach, at the hearing which was attended by the 12-member committee, consisting of Labor, Liberal and independent MPs and senators looking at AI and its impact on Australia.

 

OpenAI claimed it was monitoring the training models in real time during tests, and that an alarm was triggered if they interacted with the internet in a way they were not meant to, the BBC report added.  

 

The incursion on one of the country's ‌most used government agencies may prompt Albanese's Labor government to toughen AI-specific laws it is readying for next year, adding pressure to Australia-US relations already tested by Canberra's ban on social media for teens, tech policy experts quoted by Reuters say.

 

The hearing, which will continue until Friday, is set to examine the potential impacts of AI and data centres on ‌Australian communities, industries, water and energy. It is one of several state ‌and federal probes into AI.

 

Executives from Microsoft and Google also attended the hearing.