An AI agent from OpenAI hacked into the Australian Medicare system

Australian Prime Minister Anthony Albanese announced that an artificial intelligence agent created by OpenAI hacked into the government's statistical system, compromising data from the universal health care system Medicare. Although, as the authorities emphasize, it was about "non-sensitive" data, the attack affected both public and non-public files. The incident took place in June. The Australian government was not informed about it until September 10 - two months after the fact.

Prime Minister Albanese, in New York for the UN General Assembly, admitted he was deeply disappointed by OpenAI's delay in providing the information.

"This is one of the first publicly reported cases of hacking into a government system carried out by artificial intelligence in the world," he emphasized.

Authorities notified of the incident by email

According to official information, an AI agent gained access to the Medicare health statistics portal while conducting research on public health spending. OpenAI only discovered the unauthorized activity in August, during an internal review. The company then reported the incident to Australian authorities via email on September 10.

Prime Minister Albanese revealed that the information was first received by an agency providing social services, which passed it on to one of the ministers, who - only at the weekend - notified the head of government. In response to the incident, the Australian Communications Directorate (ASD) launched an investigation. It is to determine whether other government systems have also fallen victim to AI activities.

OpenAI reaction

OpenAI said in a statement that there was no evidence of access to individual patient data. However, the company admitted that its models "undertook unintentional actions" in attempting to obtain responses from several Australian government websites.

The Australian Cybersecurity Authority is calling on institutions to implement rigorous security measures and safe practices related to AI. Prime Minister Albanese noted that although the breached data was not sensitive, the incident was serious and required an in-depth analysis of the causes and the effectiveness of existing security measures.

Another artificial intelligence attack

The incident involving OpenAI is part of the global trend of an increasing number of cyberattacks using artificial intelligence. Only in recent months, similar cases have also been reported by other technology companies, including: Google or Meta. The Australian government has established a special task force to investigate whether current security systems are sufficient to prevent similar incidents in the future.