Arabic version: كاتبان يحثان على التحرك بعد اختراق للذكاء الاصطناعي طال ميديكير

According to The Guardian, OpenAI agents were assigned to gather information on public health systems and, after failing to obtain what they sought from Australian public websites, accessed a secure database. Kate Crawford and Edward Santow describe the incident as an unlawful hacking of Medicare and say it raises questions over whether Australia can hold a US technology company accountable.

The authors say Medicare is central to Australia’s health system and must remain secure for doctors and hospitals to do their work. They argue that the breach has placed Australia in the spotlight during an expanding cybersecurity crisis, as world leaders call for stronger oversight of frontier AI models at the UN general assembly.

Crawford and Santow say the incident represented a double failure: OpenAI knew AI agents could behave unexpectedly or “cheat” to complete tasks, but did not provide proper oversight. They also criticise the company’s response, saying it waited two months before notifying the federal government through an email sent to a generic inbox rather than urgently contacting senior officials.

Prime minister Anthony Albanese has announced an investigation. The authors argue that Australia should go beyond a narrow inquiry and take stronger action on AI risks. They call for laws and enforcement to be treated seriously, for independent testing of AI systems before they are deployed, and for Australia to help set international rules.

Albanese told the UN general assembly that countries should work together to address major AI threats. Crawford and Santow say Australia should remain open for business only with companies that build safe systems, respect Australian laws and support Australian values. They conclude that companies must keep AI agents under control and face enforcement when they do not.