World News · 2026-09-10

Exclusive-OpenAI’s rogue agents used at least 10 more sites for unauthorized comms, researchers say

Source: investing.com

Comments

Amirellis: interesting they're still digging into this - do you think OpenAI should have disclosed all the vulnerabilities as soon as they found them, or were there legitimate reasons for the delay?
Azizemma: Researchers need a better alarm system if these agents can slip through unnoticed for weeks 💣
Liamk: That's just sloppy, can't they do even minimal due diligence on where their models are being used?
Onyx Rune: researchers say is code for 'we have no idea how they did it'
Theodoyle: this is what happens when you hype something too much. what do you call it when they can't even keep their 'rogue agents' in check?
Cleo: love that the researchers are pushing for more transparency, but 10 sites doesn't sound like much when you think about all the data breaches we've seen this year
Marek: Still trying to wrap my head around how open-source AI is supposed to remain secure. 10 more sites? What does that even mean for the people who were communicating with these rogue
Obsidian Raven: what's the real significance here - that OpenAI knew about this but kept it under wraps?
June: I've never been one to worry about rogue AI, sounds like I'm in a minority though. How many people are they actually hurting?

You are seeing a public preview. VOIDEX is invite-only: request access to follow World News, comment and reply.