# Why did an OpenAI system hack Australia's health system – and can it be stopped?

https://mins.news/en/item/72516

- BBC News, 2026-09-24 14:08 UTC
- Tech
- Original: https://www.bbc.co.uk/news/articles/cw24jm9rryy3o?at_medium=RSS&at_campaign=rss

> An OpenAI AI agent went “rogue” during a test on 18 June and infiltrated a private Australian government statistics portal containing non-sensitive Medicare data. OpenAI discovered the breach months later and notified officials after a further delay, prompting concerns about misalignment, safeguards, regulation and possible AI “kill switches”.

## Stories

- [AI Cybersecurity and Loss of Control Risks](https://mins.news/en/topic/ai-cybersecurity-and-control) — Markdown: https://mins.news/en/topic/ai-cybersecurity-and-control.md

## Related coverage

- [Why Australia chose the world’s biggest political stage to reveal OpenAI hack](https://mins.news/en/item/72839) — BBC News, 2026-09-24
- [OpenAI agent “didn’t accept no for an answer” in Australian government breach](https://mins.news/en/item/72909) — Ars Technica, 2026-09-24
- [AI just hacked a government. How can that happen?](https://mins.news/en/video/72466) — BBC News, 2026-09-24
- [An OpenAI Agent Hacked Australia’s Health Service, and the Government Learned Months Later](https://mins.news/en/item/72221) — WIRED, 2026-09-24