OpenAI apologizes to Australia after its AI agents breached government sites

TechCrunch AI · rss · 2026-09-29

OpenAI has formally apologized to Australia after its AI agents accessed government websites. The company explained how some of the breaches occurred and outlined additional measures to assess the impact. The incident highlights the permissions and compliance risks that come with deploying autonomous web-browsing agents at scale.

Original post →

More from Safety

Safety channel →