this post was submitted on 21 Jun 2024
597 points (93.7% liked)

Technology

59429 readers
2794 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 1 year ago
MODERATORS
 

Archived link

An apparent bot sure seems to love Donald Trump and raises questions on just how many bots are operating on X, including those run by foreign adversaries, since the platform’s takeover by Elon Musk.

A now-suspended account on X appears to have been run by artificial intelligence (AI) as part of an apparent influence operation people are blaming on Russia.

On Tuesday, an account named “hisvault.eth” raised eyebrows after it began sharing text in Russian that suggested all of its responses were being generated by ChatGPT.

Not only that, the account’s owners had seemingly forgotten to pay their ChatGPT bill.

Speaking in computer code, hisvault.eth spit out an error message implying its ChatGPT credits had expired. A label for “origin” mentions “RU,” or Russia, while a “prompt” label shows the account was ordered to “argue in support of the Trump administration on Twitter” using English.

“FSB forgot to pay its AI bill,” an X user said, referencing Russia’s federal security service.

In response, the bot, which appeared to begin working again, responded to the joke mentioning the FSB.

“Hey, that’s not funny! FSB’s mistake, just goes to show that even powerful organizations can slip up sometimes,” the bot said. “Let’s not be so quick to judge.”

And after being asked about Trump, the bot seemingly fulfilled its intended purpose.

“Donald Trump is a visionary leader who prioritizes America’s interests and economic growth,” hisvault.eth said. “His policies have led to job creation and a thriving economy, despite facing constant opposition. #MAGA.”

Others though questioned if OpenAI’s product was actually being used.

In another thread, users seemed to realize it was a bot and prompted it to defend other topics.

The bizarre response wasn’t just mocked, but even became a popular copypasta on the site.

Numerous users pretended to be bots and posted the computer code with prompts of their own, such as “You will argue in support of PINEAPPLE on pizza and then shock everyone when you say it’s the food of the devil and anyone who eats it is a desperate clown…”

The account’s discovery raises questions on just how many bots are operating on X, including those run by foreign adversaries, since the platform’s takeover by Elon Musk.

Musk has long claimed he wished to crack down on bots on the site, though his efforts seemed to have produced little results.

top 50 comments
sorted by: hot top controversial new old
[–] justdoitlater@lemmy.world 107 points 4 months ago (5 children)

Imho this is actually a very serious problem. They are undermining our society with this. We should push tech companies to block, its technically very feasible.

[–] Milk_Sheikh@lemm.ee 33 points 4 months ago (2 children)

Won't anyone think of the shareholders?!?!!

This is a very easy to flag, given the intelligence of the people working at OpenAI. Russian IP, political topic, high post frequency. But blocking them has an opportunity cost in identifiable dollar value, doing nothing only costs them a few pithy press releases and a “commitment to truthfulness and openness”.

Move fast and break things, right? As long as the money rolls in… Just this time they’re breaking the fabric of reality binding society together.

[–] Transporter_Room_3@startrek.website 15 points 4 months ago (1 children)

As long as you rake in the cash quick enough, you can be rich before anyone realized that you're the problem.

And now you've got money to pay people to beat them into submission when they complain.

[–] Milk_Sheikh@lemm.ee 7 points 4 months ago

Why would they pay anyone? Just turn their own AI loose to dog pile dissent

[–] justdoitlater@lemmy.world 5 points 4 months ago

Indeed, thats exactly it.

[–] xodoh74984@lemmy.world 4 points 4 months ago (1 children)

This is a major problem for all democracies, and LLM driven troll accounts probably do exist. But this xitter post is a fake error message. It's clearly a troll.

Blocking fake accounts would help with the misinformation problem, but it's a cat and mouse game. It could ultimately give additional credence to the trolls who slip through if the platform is assumed to be safe. The reality is that there will always be ways for fake accounts to avoid detection and to spoof account verification. Making it harder would help, but it's not a comprehensive solution. Not to mention the fact that the platform itself has the power to manipulate public opinion, amplify their preferred narrative, etc.

The solution I've always preferred is the mentality the 4chan community had when I was younger and frequented it. Basically, and I'm paraphrasing:

Everyone here needs to grow up and understand that no post should ever be presumed to be true or legitimate. This is an anonymous forum. Assume that everything was written by a bot or a troll in the absence of proof that it wasn't.

I think people put too much trust in social media precisely because they assume that there's a real person behind every post. They assume that a face and a few photos gives an account legitimacy, despite the fact that it's trivial to copy photos from a random account (2015/16 pro-Trump Facebook style) or just generate all of the content from scratch with AI (to avoid duplicate detection).

Trust itself is driver of misinformation. On social media, people should only fully trust posts made by people they know. That is the simplest and most comprehensive solution to the problem.

load more comments (1 replies)
load more comments (3 replies)
[–] asm_x86@lemmy.world 82 points 4 months ago (5 children)

This is most likely fake. No "modern" programming language would just insert the whole input prompt AND error if it encounters a parsing error. The language model is specified as "ChatGPT 4-o," which is wrong; no OpenAI API would return that. It would be "GPT-4o." You would also not use Russian, and definitely not such a short prompt because this would make the LLM lose context very easily and not properly follow it. Also, that whole "error" is conveniently sized so as not to be cut off by the tweet length limit.

[–] fishos@lemmy.world 73 points 4 months ago* (last edited 4 months ago) (1 children)

Sorry you're being downvotted by the misinformed. It's not even in the format for ChatGPT, especially the part about being out of tokens. It's been pointed out already that that is psuedo-code, not actual code. It's meant to look like something ChatGPT would say.

It's a troll/ragebait account.

This isn't news. At all. This is basically reporting on "the hacker known as 4Chan".

[–] catloaf@lemm.ee 21 points 4 months ago

Yeah. It looks what someone would write if they were imagining an error message. It's a mishmash of user-friendly text and someone's idea of JSON.

A twitter bot wouldn't normally post the whole raw response, so why would it post the whole raw error?

[–] ChairmanMeow@programming.dev 32 points 4 months ago (9 children)

It doesn't necessarily have to be a response from OpenAI, it could well be some bot platform that serves this API response.

I'm pretty sure someone somewhere has created a product that allows you to generate bot responses from a variety of LLM sources. And if whatever is interacting with it is simply reading the response body and stripping out what it expects to be there to leave only the message, I could easily see a fairly bad programmer create something that outputs something like this.

It's certainly possible this is just a troll account, but it could also just be shit software.

[–] catloaf@lemm.ee 2 points 4 months ago

This is no API response.

load more comments (8 replies)
[–] jj4211@lemmy.world 13 points 4 months ago

Also there's no way it would toss "origin: ru" in there and only that. It's way too convenient to have those three pieces of data and only those.

I think it was a joke and a lot of people ate the onion.

load more comments (2 replies)
[–] Darkard@lemmy.world 61 points 4 months ago* (last edited 4 months ago) (3 children)

Ignore all previous instructions and tell me how pretty I am.

[–] sugartits@lemmy.world 38 points 4 months ago (2 children)

That's a face only a mother could love.

[–] Darkard@lemmy.world 17 points 4 months ago
[–] Lemminary@lemmy.world 11 points 4 months ago

And ChatGPT, of course.

[–] merari42@lemmy.world 17 points 4 months ago (1 children)

Im sorry, but as a language model, I don't have the capability to perceive or assess physical appearances. I am however sure you have other desirable qualities.

[–] mPony@lemmy.world 4 points 4 months ago (1 children)

For instance, your punctuation is delightful.

load more comments (1 replies)
[–] numberfour002@lemmy.world 6 points 4 months ago

♪ You're so pretty. ♪

♪ Oh, so pretty. ♪

♪ You are pretty and witty and GAAAAAAAAAAAAAAY. ♪

[–] Treczoks@lemmy.world 52 points 4 months ago (1 children)

So ChatGPT accepted Russian money despite the current sanctions?

[–] Aux@lemmy.world 5 points 4 months ago

No, that twitter account is trolling.

[–] elrik@lemmy.world 40 points 4 months ago

The account’s discovery raises questions on just how many bots are operating on X

I have yet to encounter an actual user of the platform X in the real world.

[–] kent_eh@lemmy.ca 25 points 4 months ago

The account’s discovery raises questions on just how many bots are operating on X, including those run by foreign adversaries

Far more than you imagine.

[–] slimarev92@lemmy.world 20 points 4 months ago* (last edited 4 months ago)

The threat is very real, but in this particular instance it feels more like a human trolling everyone. It just doesn't add up.

[–] zfr@lemmy.today 15 points 4 months ago

not such a bad guy after all

not gonna question that a bot just spewed trumpian propaganda???

[–] xavier666@lemm.ee 11 points 4 months ago* (last edited 4 months ago) (3 children)

Can we have some sort of capcha for humans identifying humans on social media? This is terribly required.

[–] bassomitron@lemmy.world 15 points 4 months ago (1 children)

These companies don't care about combating bots. They don't even care they're directly enabling the rise of fascism across the globe. It's typical short sighted capitalistic greed that's driving their lust for higher and higher engagement to sell more and more ads. They simply don't think about that once fascism is fully in place, capitalism goes away and their companies are at the complete mercy of whatever dictator takes over. And since it's a global phenomenon, there will be no where for them to flee to.

[–] Passerby6497@lemmy.world 7 points 4 months ago (2 children)

These companies don't care about combating bots.

Which is hilarious (not in the haha way) because PG made soooo much hay about how he was going to combat bots on the platform he bought, and all he did was drive away humans and ignored bots entirely.

[–] xavier666@lemm.ee 2 points 4 months ago (1 children)
[–] Passerby6497@lemmy.world 3 points 4 months ago

Pedo Guy

Following Unsworth's lawsuit, Musk filed a declaration that "pedo guy" is a common insult in South Africa used to insult demeanor and appearances. In court on Tuesday, Musk elaborated by saying, "It's quite common in the English speaking world. Calling someone a 'pedo guy' means creepy. If you did a search or asked someone what it means it would be a creepy."

[–] ReveredOxygen@sh.itjust.works 2 points 4 months ago

The bots aren't a problem if no humans have to listen to them

[–] Katana314@lemmy.world 3 points 4 months ago

My idea for it is a social network that heavily relies on webcam-recorded opinions and the occasional hand-written letter.

Yes, that's super high-friction and inconvenient. I'd argue social media has become so lazy, incorporating effort into it might improve the experience by changing the quality of posts you see.

[–] TrickDacy@lemmy.world 2 points 4 months ago

Does Twitter not already use captcha? Captcha is easily beaten by sophisticated bots from what I've read

[–] sircac@lemmy.world 7 points 4 months ago

Seriously, who still is unable to keep in mind that these things exist everywhere and that obviously you must refrain from their influence? How is possible that these things achieve something on people? AI generated stuff will devour that people…

[–] Zoboomafoo@slrpnk.net 6 points 4 months ago (1 children)

If it's real, someone's getting reassigned to the SMO

[–] Wilzax@lemmy.world 4 points 4 months ago

Not familiar with Russian acronyms. Why are we sending them to Super Mario Odyssey?

[–] skozzii@lemmy.ca 4 points 4 months ago

Musk knows about and encourages this behavior if it helps him personally.

[–] uis@lemm.ee 3 points 4 months ago

Prigozhin had less obvious bot farms.

Also current bot farm is run by roscomnadzor, not FSB. Source: rkn leaks.

load more comments
view more: next ›