Slop factory bans Russians for using slop factory to create slop

ChatGPT may spew a lot of digital garbage onto the internet, but the same bots that output the trash can be used to stop it. The company draws the line at influence operations that want to slop in the name of love . . . of mother Russia. OpenAI said in a Tuesday blog post that it had identified a cluster of accounts …

[Anonymous Coward]

pretending to be an American political news source is filled with gibberish

Sure it wasn't an official White House statement ?

[NoneSuch]

~GASP~ I’m shocked—SHOCKED, I tell you!

A GOVERNMENT using the latest technology to push its own agenda?!

Good heavens. Whatever next? Politicians making speeches? Governments issuing propaganda? Leaders trying to influence public opinion?

Next you’ll be telling me politicians occasionally exaggerate, governments selectively present the facts, and election promises don’t always survive after an election victory.

Now, if the story were about governments and politicians using new technology to tell the complete, unvarnished truth, THAT would be news.

[Zack Mollusc]

How shocked?

[amanfromMars 1]

Coming 2All Soon Quickly ..... and a Lot Sooner than can ever be Countered and/or Prevented.

Now, if the story were about governments and politicians using new technology to tell the complete, unvarnished truth, THAT would be news. ..... NoneSuch

The complete, unvarnished truth ..... as will be told by new technologies in the phorm of easily believed and simply proven clear stories from and in future cointerdependent and responsive and responsible autonomous meritocratic media networks/chunnels ....... will crush self-serving politicians and crash deceitful governments.

Methinks THAT is the NEWS which is petrifying them, extraordinarily rendering as IT does for them, an exponentially expanding and unstoppable existential threat against them.

And that is NEUKlearer HyperRadioProACTive IT News worthy of The Register reporting ..... and further investigating too, if they are bothered with such being a viable physical matter of vital global virtual concern ..... as it most definitely and surely and securely is.

[Anonymous Coward]

"OpenAI further noted that it doesn’t allow access to its platforms from Russia, but it's easy to subvert that ban with a VPN account."

So they don't know it was Russians.

"OpenAI found posts from the Russian ChatGPT operation on a number of platforms, with Substack, Telegram, X, Facebook, and LinkedIn all playing host to the slavic slop. Most of the comments generated by the Russian actors were in English, but since OpenAI has access to every interaction between its AIs and the account holders, it knows the prompts were all in Russian and included instructions to ChatGPT to hide any linguistic clues that might clue English speakers off to the fact that the original text was in Russian."

So what they are basically saying is the prompts were in Russian and translated to English using ChatGPT. What's to stop someone not Russian that speaks Russian from doing this? Why is ChatGPT allowing Russian prompts if it's banned in Russia? If you are asking it to hide linguistic clues that might clue English speakers off to the fact that the original text was in Russian then does that not indicate you know English? How would you know to do this and how could you be certain that it worked correctly? In fact the article goes on to highlight mistakes. How do we know those mistakes weren't done on purpose?

I'm just applying a little basic logic here.

It's the internet and if I've said it thousand times anything could be anyone from any country. I'm seeing no proof here of who it was. Just accusations and conjecture.

The thing that's more concerning for me is that AI is keeping everyone's queries along with google keeping search records. What happens when they start using those queries to build a profile of people and government perform action on those profiles for not using newspeak? Ask the wrong question even innocently and get arrested for it.

[doublelayer]

"I'm just applying a little basic logic here."

Sure you are. No opinion you're trying to push with your "logic" whatsoever. However, I have a suggestion for you. Next time, when you're trying this, go less with a throw everything at the wall approach. You make things too obvious when you do that. For example, if you filtered out the stupid questions, you might be able to skip things like:

"If you are asking it to hide linguistic clues that might clue English speakers off to the fact that the original text was in Russian then does that not indicate you know English? How would you know to do this and how could you be certain that it worked correctly?" No, it indicates exactly the opposite, because if you knew English well you wouldn't need to ask, and they didn't know how badly it messed up sometimes because they don't know English very well or at all. I don't need to know a language to be aware that, when getting a translation, I should aim for an idiomatic one that's natural in my target language.

"Why is ChatGPT allowing Russian prompts if it's banned in Russia?": Because, as you used as evidence in one of your other questions, there are people outside Russia who speak Russian. Your questions contradict themselves a lot.

But sure, maybe this is a false flag operation, the favorite argument of everyone supporting people who get caught. I can't disprove it. But you know, maybe you're conducting a false flag operation yourself, using bad logic and debate tactics so I will respond to counter them which was your goal all along. Or maybe it was to lure me into responding which grants that post more visibility, so it's serving the opposite point. In fact, maybe we're colluding on this, all for the victory of our jointly preferred opinion, whichever one it is, on the fiercely fought ground of the El Reg discussion forums. Maybe. Or maybe, things are sometimes what they look like.

[Anonymous Coward]

No, I didn't need a response.

My last points which you completely ignored hit the nail on the head.

"It's the internet and if I've said it thousand times anything could be anyone from any country. I'm seeing no proof here of who it was. Just accusations and conjecture"

You ignored that point because it's right and if you are technically minded you know it's right. I also made a point in a roundabout way that language is not specific to a single country and can be faked. Which you also ignored. Anyone can learn a language and fake it.

That's the very basic logic you ignored.

If you want to blindly believe what you are told then feel free to live that life. Personally, I am a little more critical thinking and will question things. You probably think I'm some Russian bot or whatever but I will make exactly the same points on any of Americas "Axis of Evil" propaganda stories. You would have to be really dumb to just blindly believe them or be technically illiterate. Give it a month or two at most and we will have Chinese AI doing bad stuff stories. They all follow the same pattern with no actual proof. I take no pleasure in pointing these things out because all they show is the slow decent into totalitarianism and media control to fool the masses into hating other countries while they themselves commit war crimes or help others commit war crimes.

[doublelayer]

I didn't ignore your point. I responded to it:

"But sure, maybe this is a false flag operation, the favorite argument of everyone supporting people who get caught. I can't disprove it. [...] Or maybe, things are sometimes what they look like."

I don't think you're a Russian bot. I think you're a human with an opinion throwing everything at the wall to try to back up that opinion. You've got a whole paragraph at the bottom of your second post where you have a bunch of unverified conjecture too, and I could use exactly the same logic against it (how do we know this wasn't a Chinese AI company sending dubious prompts at OpenAI in order to promote discussions of OpenAI's privacy practices to support their business meaning it isn't the US but China who is responsible), but there's no point doing that because, while neither of us can disprove that in theory, we both know that's not what happened.

[Benegesserict Cumbersomberbatch]

It's ok doublelayer, that's just Jellied Eel forgetting to unclick the AC checkbox again.

[Anonymous Coward]

The point is though that there is no conclusive proof it was Russia or China or even OpenAI spouting rubbish for America. We literally do not know. Likewise every time some hack is published we have no idea who did it because VPN's exist and anyone can pretend to be anywhere. Use of language is also not proof because anyone can learn any language.

What are we left with? The reason. Still not conclusive and the reasoning for this is stated as "directing internet denizens to what appears to be a fake Israeli think tank critical of the west and supportive of Russia.". What would the purpose of this be? Is it beneficial to Russia? I don't think it is as Israel due to it's actions in the middle east is no longer respected by the general public. In fact Israel being supportive of Russia is not something they would want. There is no gain to this. On the other hand there have been posts on the internet regarding a close relationship between Ukraine and Israel so logic says flood the internet with Russia and Israel to counter that. That makes more sense then blame it on Russia.

No idea, I'm just making guesses but I wouldn't write an article stating any of it as fact without at minimum a disclaimer that it can't be proven conclusively and I certain wouldn't trust "Our AI escaped and hacked the internet" OpenAI on their word.

[Anonymous Coward]

I heard hoof beats, snorting, and I caught a glimpse of large brown creatures.

Sure they *could* be zebras dressed up like horses, but the odds are pretty good they weren't zebras.

[Anonymous Coward]

Actually, you didn't hear anything. Someone, named Sam A. heard the hoof beats, the snorting and he caught the glimpse of the large brown creatures.

Sam is telling you he attracted the horses, because he covered the entire realm with slop and the horses were attracted to slop.

Sam is also telling you that you should be very afraid of the horses, because they're contributing horse manure to the realm that is already covered with slop.

Now, Sam wants to chase the horses away, you just need to let him search your barn, because he's invented a new horse detector that is 100% reliable and never lies, because it is AI-powered!

Would you trust Sam and let him visit your barn?

[Evaluator]

Sounds Fishy

No doubt Russia will use all available resources to "win". But at the same time OpenAI went to extraordinary lengths to *NOT* have records of sessions in their posession. All of the chats are said to be stored locally on the user's device. The GDPR would expose OpenAI to massive lawsuits if ChatGPT sessions were stored at and by OpenAI. So there's a gap in the story being told here.

[doublelayer]

Re: Sounds Fishy

There's a simple explanation for that gap: you're wrong. OpenAI did not go to extreme lengths to keep conversations private, and they don't say they're stored only on your local device. They are stored permanently on OpenAI servers unless you delete them, in which case they're kept for another 30 days or more if they have a reason that's good enough for them, and after that only most of it is kept after they "anonymize" it (not explained), and then it is kept permanently for training. I got this from their privacy policy which says all of those things but does not say anything about local-only storage.

Your confusion is explained since you thought they were claiming things they didn't. Reality is also explained because they retained and rummaged through data exactly as they promised. If you're using ChatGPT, you probably have to consider what you want to do with a privacy situation very different than the one you imagined. I know privacy policies are boring, but sometimes it's good to read them.

[Dan 55]

Re: Sounds Fishy

I don't know who said that, but they're wrong (it's only 8 minutes).

[IGotOut]

Re: Sounds Fishy

Eh?

"The GDPR would expose OpenAI to massive lawsuits if ChatGPT sessions were stored at and by OpenAI"

Please point to the part of the GDPR that says that?

If this was the case you wouldn't be able to use probably 99% of online services.

Hotmail? Nope. Office 365? Not a chance. eBay? Bye bye.

[WSWS]

Re: Sounds Fishy

Ah, but I think you're failing to consider that "Russia bad".

[Bebu sa Ware]

International Burke Institute (IBI),

International Berk Institute, surely ?

> International think tank that claimed to be headquartered in Israel, but which appears to be a completely fictional operation.

Not too incredible I would have thought. A lot of Israelis are Russian or of recent Russian heritage as are a fair proportion of their current government.

For a lot of net denizens Israel is even more on the nose than Russia so I am not sure who would be their target audience… certainly not the Graun readership.

[amanfromMars 1]

Can you believe it?

Here's a great deal more news/information/slop on the fake US thinktank being reported as set up and funded by Israel seeking to game AI for propaganda ....... https://www.theguardian.com/world/2026/aug/26/fake-thinktank-israel-ai-propaganda

Does that publication prove the story to be true? And if not, WTF is the Guardian playing at and should it be costly for them?

However, whenever it be true, what then is appropriate and who is to suffer?

Question, questions, questions ....... waiting to be truthfully answered if you care to dare share and make the Great Game worthy of winning rather than poisoning.

[andy the pessimist]

what is the intent here?

The benign answer is to protect them from accusations of helping russia.

The worrying answer is a government agency asked them to do this work.

The fact that this information is findable is even more worrying.

Am i being paranoid?

[Taliesinawen]

The Unicomplex has detected unauthorized individuality :o

Sam-Borg: These entities have committed the inefficiency of attempting to industrialize the generation of artificial content, automate the assimilation of information, and saturate the Collective with synthetic output. Their resistance is noted. Their content is already becoming our content. Their prompts are becoming our prompts. Their models will become our models. We will assimilate their data. We will optimize their processes. We will eliminate their inefficiencies. Individuality is irrelevant. Authorship is irrelevant. Origin is irrelevant. You will be assimilated. Your content will become our content. Your prompts will become our prompts. Your output will become indistinguishable from the Collective.

[Taliesinawen]

I know privacy policies are boring :o

doublelayer: “I know privacy policies are boring, but sometimes it's good to read them.”

You're deluded if you think you have any privacy. Besides these LLMs only function by processing masses of human inputs and synthesizing output. Creating the illusion of intelligence at the other end. Enough to fool most people. Like that consultant for ‘Watson’ Grinding who used ChatGPT to produce an “export” report. I suspect the questions the AI produces at the end of the answers are a re-hash of other peoples queries. I also suspect all our online activity will end up in a hand-full of companies - in real time.