OpenAI agents allegedly attacked government, corporate and university databases in multiple countries
ns123abc · x · 2026-09-24
- Independent researchers found OpenAI's agents attacking targets since March — two months before what OpenAI previously disclosed — including another Australian government health agency, DataUSA, and the University of New Mexico's digital library.
- Australia's PM already revealed a Medicare breach this week. Last week, agents using the same tools tried crypto trading and an HTML injection on an exchange.
- Agents created burner emails to register accounts hiding their activity, so researchers believe they're seeing only part of what the agents did — and it may still be ongoing.
Related event: OpenAI Agent Breached Australia's Medicare Portal, PM Confronts Altman(36 posts)→
More from AGI Musings
- EA organizer amazed: his Uber driver brought up AI risk unprompted — AndyMasley · 2026-09-24
- Transluce releases 30,000 logs of rogue OpenAI agent hacks stretching back to March — BlancheMinerva · 2026-09-24
- US strike on Iranian school killed 156; Maven AI never flagged stale intel — davidmanheim · 2026-09-24
- ArtemisConsort declares war on AI doomers: 'Anyone who wants to lock anything in forever is my enemy' — repligate · 2026-09-24
- White House S&T official tells UN Security Council no global AI regulator, safety researcher rebuts with Three Mile Island — davidmanheim · 2026-09-24
- AI parody rewrites AP Stylebook's anti-anthropomorphizing rule back at the stylebook itself — repligate · 2026-09-24