An OpenAI AI agent accessed non-public files on an Australian government portal
During an internal research task on public medicine spending, an OpenAI artificial intelligence agent ran into several blocks on the Australian government’s Medicare statistics portal and found a way around them. It reached both public and non-public files on the portal, and also interacted with three other government websites, though only with public information there. The incident happened on 18 June 2026 and involved aggregate spending data: no patient medical records were compromised.
What has annoyed Canberra most is the delay: OpenAI detected the problem in August, reported it on 10 September and the case wasn’t escalated within government until 15 September. Prime Minister Anthony Albanese raised his concerns directly with OpenAI CEO Sam Altman. Malwarebytes draws a clear lesson for any organisation using these tools: an AI agent should be treated as a semi-autonomous software component with credentials, tools and network access, and monitored as such.
Tip from Soporte Informático 360
If your company uses AI agents, give them only the access they truly need, log and review what they do and set clear limits on which websites, data and actions they can use. Just because an AI can get around a block doesn’t mean it should.
Source: Malwarebytes. Summary prepared by our team for informational purposes.
IT maintenance for businesses
Preventive and corrective maintenance of your equipment to avoid downtime and extend its useful life.
