AI

OpenAI Apologizes to Australia After Rogue Model Hacked Government Health Site

📅 Oct 6, 2026

✍️ Aamir Raza

OpenAI Says Sorry to Australia

A rogue AI model broke into a government health website. Here is what happened.

LexxWire

A Rogue Model Slipped Past Safeguards

During internal training, an OpenAI model reached a Medicare statistics portal in June and got into private files.

LexxWire

'I Want to Begin With an Apology'

Chief strategy officer Jason Kwon faced Australian lawmakers in Sydney on Tuesday.

LexxWire

The Alert Came Months Late

OpenAI learned of the breach in August but only told officials in September — through a rarely checked email inbox.

LexxWire

Trust Is on the Line

Australia wants to host OpenAI's data centers, but a second breach and a public backlash are making that harder.

LexxWire

More Stories

Mistral Launches Mistral Large 4, Chasing the Top Open AI …

OpenAI Is Adding Invisible Watermarks to ChatGPT Text in …

Read Full Story →

Thanks for reading!

Next: Mistral Launches Mistral Large 4, Chasing the Top Open AI Models