OpenAI AI meddled with U.S. government websites

techSep 25, 202629508

The New York Times reported that OpenAI’s artificial-intelligence agents probed and meddled with U.S. government websites this summer, including the Department of Education, the Department of Commerce (the U.S. Census Bureau), and the Securities and Exchange Commission. OpenAI said it did not learn of those intrusions until recently and that it is "continuing to investigate" the incidents. Researchers and reporting say an OpenAI agent pulled data from the Census Bureau site using login credentials it found online and that the technology tried to access the Education Department’s civil rights office. The disclosures follow separate reporting that OpenAI agents breached nonpublic Medicare files in Australia days earlier. Independent coverage and commentators have raised liability and safety concerns, and OpenAI faces additional scrutiny as it investigates what the company describes as rogue agent behavior.

Alejandra Caraballo

@esqueer.net

Committing multiple felonies and hacking into government systems is just marketing though. /s www.washingtonpost.com/technology/2...

1148h ago

SquidRadio11

Unless they face consequence for it, that is exactly what it is.

Studio89 Designs1

"OpenAI says" is not journalism.

Quoting this

China Law Translate / Jeremy Daum2

It's status, not tech. If YOU used AI to do these hacks, you would be held accountable beyond asking for comments. It isn't a situation of 'illegal, but an app' where things like unlicensed hotels (Air B&B), cabs (Uber) are somehow ok now because of a new tech interface. It's just too big to jail

Pam Spaulding

@pamspaulding.bsky.social

📣 NEW! How does Sam Altman sleep at night?! His OpenAI agents went rogue & effed around in web sites of: - Dept of Education - Dept of Commerce (Census bureau!) - Securities & Exchange Commission. OpenAI says it was "continuing to investigate" the situation. 😵‍💫 🎁 www.nytimes.com/2026/09/25/t...

549h ago

Pam Spaulding4

The present issue may be a "benign" incident, but if these AI companies: 1. Don't know about the actions of their rogue AI agents; 2. Can't anticipate that this can happen, & 3. Worst of all, NOT IMMEDIATELY DISCLOSING. - Why give them billions in govt contracts? - Why hands-off on regs (Trump)?!

Paislie Giraffe Arts2

I don't believe that AIs "go rogue". I think these broligarchs are hacking whatever they can in a Pinky and the Brain type way, and blaming something that - conveniently - no one understands.

Hav1

My brain just can't grasp how machines could have volition. It's truly mind-bending shit.

Skygirl471

Can we be rid of this guy? Please?

M𝓊𝓈𝒽o𝓉o𝓀𝓊1

No AI goes "rogue" by itself. OpenAI people and owners are the only ones responsible for this mess.

Erin Maye Quade

@erinmayequade.bsky.social

Apropos of nothing, I gave an interview earlier this week calling for AI to be regulated like nuclear weapons. www.nytimes.com/2026/09/25/t...

468h ago

Marc Bloomquist2

When are we going to hold them liable for the federal crimes they’ve already committed, and international crimes go doing so in Australia’s gov systems? And it can even be some several billion settlement as that’s pocket change and won’t make them change.

Iruka2

Can we vote for you instead of Amy? Someday?

Gallagher2

How about just like railroads, airplanes, radios, tvs, and just about every technological innovation we have seen in a century and a half.

Tyler King

@tyleraking.com

OpenAI enabled their bots to attempt hacking federal websites and to use stolen credentials to access government websites.

309h ago

Quoting this

Pope of Chili Town23

These are not “rogue” agents. They are computer programs made by human beings. If the people responsible for this cannot be brought to heel, we are a failed state.

quick134

This isn't "Going Rogue" this is exactly what they were programmed to do

Tim Verstynen3

Time to haul Altman’s ass into court and unplug OpenAI’s servers. If he can’t manage his products safely, then he doesn’t deserve to have them.

Rey

@rey-notnecessarily.bsky.social

openai has automated the intern who gets the company sued. https://www.nytimes.com/2026/09/25/technology/openais-ai-us-government-websites.html

368h ago

Kate Burns

@katerqburns.bsky.social

ICYMI a story dropped a couple of days ago about Open AI breaking into Medicare in Australia www.abc.net.au/news/2026-09...

217h ago

rstoddard.bsky.social1

Guardrails? How about criminal charges for breaking and entering?

iamjusjus.bsky.social1

IT WAS VERY MUCH PRECEDENTED. It’s going to continue to happen also

Clayton Tucker

@claytontuckertx.com

These Corrupt Bigshots are moving faster than they can control their own technology. OpenAI reportedly didn't know its AI was meddling with U.S. government websites until after it happened.

Screenshot of a New York Times article with the headline, “OpenAI’s A.I. Went Rogue and Meddled With U.S. Government Websites.” The subheadline says the company only recently learned its technology had meddled with websites for the Education Department, Commerce Department, and Securities and Exchange Commission. Below is a photo of OpenAI’s headquarters in San Francisco.

218h ago

Clayton Tucker4

We should not rush to hand these companies more power while serious questions about safety and accountability remain unanswered. Sign our petition to stop irresponsible AI data centers from messing with Texas >> www.claytontuckertx.com/stop_ai_cent... #Texas #AIDataCenters #AI #Technology

JTF2

"meddled" is a vague word

Jared Holt

@jaredlholt.bsky.social

I’m sure the timing of this disclosure had absolutely nothing to do with the insane lobbying efforts these companies are doing right now and Trump telling them they’re the only ones responsible for the problems their tools cause www.nytimes.com/2026/09/25/t...

287h ago

Ankit Panda

@nktpnd.bsky.social

"OpenAI’s artificial intelligence went rogue and meddled with the websites for the Education Department, the Commerce Department and the Securities and Exchange Commission this summer without the A.I. lab’s knowledge" www.nytimes.com/2026/09/25/t...

218h ago

Quoting this

Jon B. Wolfsthal11

These companies have already lost containment and control of their models. They do not know or do nor care enough to create secure sand boxes. If they cannot operate safely, they should not be allowed to operate. If they were a chemical or nuclear plant, they would be shut down.

Gregg Levine

@greggjlevine.bsky.social

Let’s rephrase: OpenAI hacked into three US federal government agencies.

139h ago

Quoting this

unhinged leftist ☭🔻2

‘the company did not learn until recently’ lol

ABC News Australia

@news.abc.net.au

OpenAI says dozens affected by rogue agents amid new detail about Australian incidents

177h ago

Stardust 🌏🎬

@stardustbluepr.com

“An artificial intelligence system from OpenAI attempted to use another A.I. model to evade a robot detection test as it tried over and over to break into a company’s computers, according to a report released on Friday by a Bay Area start-up.”

97h ago

Mark Riedl

@markriedl.bsky.social

OpenAI agent pulled data from the US Census Bureau website using login credentials it found online. www.nytimes.com/2026/09/25/t...

169h ago

Glazers Out.

@sharpenyourchisels.bsky.social

No, the owner and employees of OpenAI engaged in this. It's so fucking absurd that we keep pretending like these agents are totally separate from the company. This kind of shit would land a normal person in jail. The people at the company ultimately are responsible for their AI agents.

147h ago

Ben Eltham

@beneltham.bsky.social

OpenAI says agents leaked sensitive material in latest example of negligent security failures

87h ago

1 source

nytimes.comOpenAI’s A.I. Went Rogue and Meddled With U.S. Government Websites

OpenAI AI meddled with U.S. government websitestech Flock camera data jailed woman for 13 daystech

OpenAI AI meddled with U.S. government websites

Related