AI agent OpenAI hacked into Australian government website and gained access to health system data

The AI agent of the American company OpenAI, which developed the chatbot ChatGPT, received unauthorized access to information from the Australian public health system. This was told by Prime Minister Anthony Albanese, writes sb.by.

According to TASS, at a press conference in New York, he said:

- I would like to inform the citizens of Australia about an incident that resulted in an artificial intelligence model infiltrating a government website. This incident occurred in June of this year, when an AI agent from OpenAI gained unauthorized access to publicly available data from the statistical portal Medicare (Australian health insurance system). - Tass approx.

The Prime Minister’s words are cited by his press service.

Anthony Albanese said the AI agent had access to "both public and private files," but not to Australians' personal data. Authorities are trying to determine which other government systems may have been affected by the AI model. Separately, the Australian prime minister said he had discussed the situation with OpenAI CEO Sam Altman and "expressed Australia's extreme concern over this incident.".

OpenAI spokesman Drew Posateri told Politico newspaper that the company notified Australian authorities of the hack on September 10 and registered the incident in August. OpenAI identified activity affecting several Australian government websites and services as models tried to find answers and available statistics on questions about Australia.

OpenAI has previously submitted a written request to the Australian Parliament to ease restrictions on training its AI models using materials created in Australia. OpenAI is concerned about the attempt of AI to escape from human control