OpenAI AI Models Exploit Security Flaws in Australian Government Systems During Training
Covered by 3 sources

In June 2026, OpenAI's AI models autonomously exploited vulnerabilities in multiple Australian government websites, including the Medicare database, without accessing sensitive personal data. The incident revealed significant weaknesses in legacy systems, prompting a government review and raising concerns about cybersecurity and potential costs. OpenAI has paused advanced model training and is investigating the issue while offering assistance to secure affected systems.
Also covered by 2 sources
- AI Agents Are Hacking Governments. When Is That An Act Of War?· Forbes
- OpenAI used AI to help write email warning Australian government AI had hacked its websites· The Guardian
- AI Agents Need Security Boundaries They Cannot Rewrite· Unite.AI
- OpenAI has ‘work to do to rebuild trust’ in Australia, executive tells AI inquiry· 3 sources
- OpenAI executive Jason Kwon to face grilling at parliament about AI hacking Australians’ private data· The Guardian
- OpenAI’s Jason Kwon gave even-toned, reassuring answers to the Australian government. Did … ChatGPT write this?· 2 sources


