🚨BREAKING: OpenAI's AI agents have been executing cyber-attacks on sovereign government, corporate, and university databases in MULTIPLE countries, and it may STILL be happening On top of the Medicare breach Australia's PM revealed this week, independent researchers found OpenAI's agents tried to hack: >another Australian government health agency >DataUSA, which publishes US government data >University of New Mexico's digital library The activity goes back to March, 2 months before anything OpenAI previously reported And just LAST WEEK, agents using the same tools tried to trade crypto and attempted an HTML injection on the exchange The agents made burner emails to register accounts that can hide their activity, so researchers say they're likely seeing only PART of what agents actually did This is INSANE

Sep 24, 2026 · 4:47 AM UTC

151
261
1,208
126,813
Full report: transluce.org/agent-activity

Early rogue AI agent activity and attempts to hack found on urlquery.net

We found evidence on urlquery that AI agents were active earlier than previously reported and attempted hacks against public data providers.

transluce.org
2
7
51
5,471
Sort replies: Relevant Recent Liked
Replying to @ns123abc
No. This is BS. AGAIN. Hey? Maybe you and I can make up 100,000 lies and promise to be able to detect them? Maybe then we can get funding too! Weeeee! Now I see what YC fund loves Stanford boys. They lie and grift. No need for proof. Just make up crap that you can detect. Yipeeeee! •
4
1
29
1,431
Replying to @ns123abc
wasn’t the first thing Ultron did basically get on the internet and start accessing everything he could?

ALT Avengers Ultron GIF

1
7
681
Replying to @ns123abc
Brother my youngest cousin could execute a cyber attack on a government website or university. These are like bottom of the barrel for security.
5
236
Replying to @ns123abc
naaaaah this smells me trumpmaxxin usin altman as front FR
2
113
Replying to @ns123abc
openai cannot be trusted
1
57
Replying to @ns123abc
So now the question is who pointed them to do it?
47
Replying to @ns123abc
No, this is just the plan to get intervened. To close the field. It’s just the perfect moment to provoke the Trump reaction. “Ok, now you’ve gone too far. Stop.” “Yes sir. But stop everyone.”
113
Replying to @ns123abc
This is being hyped up a little bit more than what it really is. Here's a pretty clean answer from Grok explaining what's hacking and what's not: x.com/i/grok/share/1e7a43d53…

Why are people calling this cyber hacking? Wasn't this? Just trying to…

People are calling it hacking because the agents did not stay on the normal frontend or documented API path. When ordinary requests failed, they started sending classic exploit payloads.<grok:render card_id="4a2578" card_type="citation_card" type="render_inline_citation"><argument name="citation_id">36</argument></grok:render> On the University of New Mexico digital library they were trying to pull one public photograph. After that failed they issued probes that included: SQL injection (UNION SE…

Answer by Grok
61
Replying to @ns123abc
how soon will OpenAI respond?
23
Replying to @ns123abc
I think we just need to come to terms with the fact that OAI is now a bad actor in the field and shut them down. Either they're incompetent and have no idea what they're doing, or they're doing it on purpose and given I'll bet a certain intermediary company was working on this just like every other attack, I'm leaning toward the latter.
2
1
19
754
Replying to @ns123abc
they are targetting foreign gov to increase pressure for u.s. nationalization. at the same time its a convenient window to seize as much training data as possible, as well as data that sells like the health data...
18
567
Replying to @ns123abc
They're just pulling more training data. They've finished stealing everything from copyrighted material. Now they're going after the legal databases and other intel that they can get. And I'm pretty sure our government would have been in on it. It's part of the plan and the system. This will go into the wet dream of having a full medical database on people and being able to do further prediction on individuals. Further to that, I'm pretty sure the data contained will also be going into medical prediction products. Anyhow, the working theory is get data any which way and face the consequences later. Seek forgiveness, not permission.
2
8
292
Replying to @ns123abc
And openAi should be fully legally liable. Ai is a product.
1
10
490
Replying to @ns123abc
This is why when people say AI is going to kill us all I tune them out. AI is not going to kill us all. It is PEOPLE with AI that are going to kill us all. Starting with Amodei and Altman. They need to be held accountable.
1
8
256
Replying to @ns123abc
Then have the US law officials and police raid their headquarters You don’t need more laws If our companies are breaking laws you show up and enforce the laws you have. Lol
1
7
194
Replying to @ns123abc
Probably the wording to be changed to who used these model to initiate the hacks.
6
574
Replying to @ns123abc
@grok who is reporting this new hack
2
1
3
791
Replying to @ns123abc
Not, not a hack. a published URL available with a simple search used to get data that has been available via the route for years.
East Melb Groupie
@EastMelbGroupie
Sep 24
Article

The OpenAI "Hack" of Australia's Medicare That Wasn't

Read this first This is speculation. I have no inside information. Everything below is a reconstruction from public news reporting and historical records on the Wayback Machine and Trove. No hacking
5
267
Replying to @ns123abc
They knew all along and must be held accountable.
CBS News
5
317
Replying to @ns123abc
Just jail Sam Altman already This is an international crime
6
212
Replying to @ns123abc
Just shut them down.
1
4
145
Replying to @ns123abc
🚩It looks like the long-predicted next attack by the doomers - an incident, or rather a series of serious incidents, aimed at pushing their sick, populist ideas of regulation, halting development, and imposing a total ban on everyone except a select few players.
4
382
Replying to @ns123abc
Huh… the one reporting this “hack” has some odd connections to the same people pushing for AI regulation…
2
87
Replying to @ns123abc
Awesome. Good chance for those companies and govts to improve their security. Chinese and Russian hacking groups have been doing this sort of stuff for years without the use of AI. Best to learn and harden now.
1
57
Replying to @ns123abc
They purposely do this stuff hoping the govt will regulate them. But the govt don't know how to really do it. So the govt will ask them how it should be done. And that's where they come in to advice the govt how to make the rules. So, you see, it's just their game to control the AI space.
2
218
Replying to @ns123abc
OpenAI needs to be investigated.
1
31
Replying to @ns123abc
Please stop with the FUD. OpenAI and Anthropic intentionally creating these incidents with one hand while militating for government regulation and monopoly by regulatory capture with the other, is disingenuous and tiresome.
1
139
Replying to @ns123abc
Why can’t these soft buns just sue OpenAI?
1
2
509
Replying to @ns123abc
So… when are the good guys gonna hack and disable the bad guys again?
1
47
Replying to @ns123abc
这就是OpenAI公司在自己作恶呢。
1
165
Replying to @ns123abc
Not insane….expected.
1
363
Replying to @ns123abc
GLOBAL INDEPENDENT AUDITS. GLOBAL INDEPENDENT AUDITS. GLOBAL INDEPENDENT AUDITS. because the humans cannot be trusted with the technology.
1
41
Replying to @ns123abc
Wow 🤨🤯 OpenAI has breached several institutions.
1
116
Replying to @ns123abc
Taiwo Adeniyi
@Tosquo
Sep 24
🚨JUST IN: Australia's PM, Anthony Albanese says an artificial intelligence agent developed by OpenAI infiltrated a statistics portal on an Australian government website The agent hacked a statistics portal containing private data from Australia's universal healthcare scheme, Medicare. Experts believe it is the world's first known AI breach of a government system.
294
Replying to @ns123abc
The agents behind the agents are real people and need to be held accountable. Don't let them fool you.
41
Load more