AI Is Poisoning Reddit To Promote Products and Game Google With 'Parasite seo' (404media.co) 48
An anonymous reader shares a report: For years, people who have found Google search frustrating have been adding "Reddit" to the end of their search queries. This practice is so common that Google even acknowledged the phenomenon in a post announcing that it will be scraping Reddit posts to train its AI. And so, naturally, there are now services that will poison Reddit threads with AI-generated posts designed to promote products.
A service called ReplyGuy advertises itself as "the AI that plugs your product on Reddit" and which automatically "mentions your product in conversations naturally." Examples on the site show two different Redditors being controlled by AI posting plugs for a text-to-voice product called "AnySpeech" and a bot writing a long comment about a debt consolidation program called Debt Freedom Now. A video demo shows a dashboard where a user adds the name of their company and URL they want to direct users to. It then auto-suggests keywords that "help the bot know what types of subreddits and tweets to look for and when to respond."
Moments later, the dashboard shows how Reply Guy is "already in the responses" of the comments section of different Reddit posts. "Many of our responses will get lots of upvotes and will be well-liked." The creator of the company, Alexander Belogubov, has also posted screenshots of other bot-controlled accounts responding all over Reddit. Begolubov has another startup called "Stealth Marketing" that also seeks to manipulate the platform by promising to "turn Reddit into a steady stream of customers for your startup."
A service called ReplyGuy advertises itself as "the AI that plugs your product on Reddit" and which automatically "mentions your product in conversations naturally." Examples on the site show two different Redditors being controlled by AI posting plugs for a text-to-voice product called "AnySpeech" and a bot writing a long comment about a debt consolidation program called Debt Freedom Now. A video demo shows a dashboard where a user adds the name of their company and URL they want to direct users to. It then auto-suggests keywords that "help the bot know what types of subreddits and tweets to look for and when to respond."
Moments later, the dashboard shows how Reply Guy is "already in the responses" of the comments section of different Reddit posts. "Many of our responses will get lots of upvotes and will be well-liked." The creator of the company, Alexander Belogubov, has also posted screenshots of other bot-controlled accounts responding all over Reddit. Begolubov has another startup called "Stealth Marketing" that also seeks to manipulate the platform by promising to "turn Reddit into a steady stream of customers for your startup."
Poisoning Reddit (Score:4, Funny)
Re: (Score:2)
They said, trying to provoke fear, "This will change the Internet as we know it!"
And I said, "GOOD! There's too many fucking hitlers on there"
Yet another reason to avoid reddit (Score:2)
not on reddit.. (Score:4, Insightful)
Re: (Score:2)
Re: (Score:2)
Nice tip, thanks, I'll try it!
Re: (Score:3)
The reason is two fold. First, with Yandex being Russian, they want every political story about the West out there for people to read. They probably even promote the stories to higher rankings so they are seen. However, try putting in a query about the thoudands of Russian losses in Ukraine or Russian ships being sunk or the oil refineries in Russia being attacked or the Belgorod region being shelled by Russian forces and see what
Re: (Score:2)
Re: (Score:2)
On my Firefox Nightly I have an extension to set Yandex as an allowed default search engine. Makes it easier to pop up another browser for a session.
Somebody fix that extension to work on mobile, eh?
Re: (Score:2)
Not on reddit, but anymore Google searches are mostly useless.
Me: "what happens when milk turns sour?"
Google's first five results: "WHY DOES MILK TURN SOUR? Section 1: Milk is a fascinating beverage that is drunk by every single mammal. Did you know ...? Section 2: we love drinking nice cold milk on a warm day with cookies! Here's more about ... Section 3: milk is produced ... Section 4: You have to be careful about milk going sour; FDA recommends ... Section 5: Why does milk turn sour? ..."
Only in the final section, after pages and pages of inane self-evident drivel,
Re: (Score:2)
...pages and pages of inane self-evident drivel...
I know this is going to go against the Zeitgeist here, but I asked Copilot (the one in Bing) "why does milk turn sour" and got a response of about half the page, with no ads and very little filler. It also matched what I knew (lactose, bacteria, lactic acid, protein folding, whey, lactobacillus etc). There were no hallucinations I could see, and it also added the links it got the data from. Much better than Google search, at least in this case.
Re: (Score:2)
I tried the same, but for a technical subject. Copilot entirely made up bullshit that didn't exist. When I informed it as such, it did the usual bot apology, and then literally repeated the same made up bullshit.
For things it can essentially "search" wikipedia for, its fine. For anything else, its dogshit bad just the same as everything else.
Re: (Score:2)
Re: (Score:2)
Re: (Score:3)
Write an algorithm that detects it (tricky but doable)
Definitely tricky - current "AI detection" is IMO inconsistent-to-dogshit in terms of accuracy, then there is issue of counter-measures being useds to train smarter systems that try to beat them.
Re: (Score:2)
It would be funny if Microsoft, despite their connection with CoPilot and OpenAI if they programmed Bing to do just that to get a leg up over Google
Seriously Microsoft, consider it, Google search is on the backfoot, now is the time to strike (they wont because they are terribly myopic in regards to AI)
Re: (Score:2)
Google have the ability to wipe out AI-generated content. Write an algorithm that detects it (tricky but doable)
We're going to be in a never-ending war. Each time we create an algorithm to detect it, they'll just get better. The same war has been raging on with CAPTCHA for years. At some point humans and AI are going to be indistinguishable. It WILL happen. Then what?
The solution is going to be to figure out how to verify that new accounts are indeed humans, that requires some form of identity verification. This has privacy implications and will not be easy.
Re: (Score:2)
Write an algorithm that detects it (tricky but doable),
How to say, "I don't understand the first thing about AI, really," without saying, "I don't understand the first thing about AI, really."
Re: (Score:2)
https://www.timeshighereducati... [timeshighereducation.com]
Re: (Score:2)
Re: (Score:2)
Right now, AI detection bots misidentify English language text from neurodivergent people and people using English as a second language as AI. I suspect it's much worse in other languages -- too many Anglophones pretend that English is the only language worth supporting.
Reddit has some supremely cringeworthy plugs.. (Score:2)
So many posts are clearly marketing campaigns, not sure that AI is needed to make it much worse.
degradation (Score:3)
Re: (Score:3)
The presumption that throwing more data and more processing power at LLMs will result in subjectively better performance has little data to back it up. I'm intrigued as to when the tipping point is reached and the valley of despair looms for this technology.
Re: (Score:2)
The presumption that throwing more data and more processing power at LLMs will result in subjectively better performance ..
One thing that is true is it will result in greater Cost; likely with a low ratio of performance improvement to compensate for the added cost of just throwing more and more resources at it.
The improvements aren't that much.. The world seemingly has lost its mind in terms of over-investing hastily in a small number of technologies.
Unfortunately it seems like Detection and Spam preve
Re: (Score:2)
Re: (Score:2)
Good. I very much would like search results free from Reddit, Quora, or Pinterest.
This (Score:2)
This is how M5 and Mycroft start killing each other >:-(
Here's the deal (Score:1)
if a youtube or reddit post mentions an amazing financial, or spiritual, or etc. advisor quickly in response to someone's story. And the story has too many upvotes in too short a time, I recognize it as spam.
IMHO, with spam like that, you go after the cloud of accounts upvoting it. Track their behavior, see if they are posting, see if they regularly vote for spam. Then shadowban or kill the accounts (let them upvote but don't show the upvotes). The advertiser can create *an* account quickly. But th
"Parasite seo" = repeating one's self. (Score:3)
seo has been a parasite since inception, and as it penetrates the web deeper and deeper, more and more used to be informational sites fall to its stupidity.
The one question this story brings up for me is just how bad does Google intend to make its results? I mean, it's already on the cusp of completely useless. How much worse does it need to get?
Re: (Score:2)
As long as it's your only option, the results will continue to get worse. Let me reemphasize that for the low reading comprehension folks: As long as you keep using advertiser driven crap, the enshitification will continue. Your opinion isn't wanted nor considered.
It will take a complete hollowing out of Google's bank account into the red to get them to stop. Given how many fingers they have in various industries, that's not going to happen short of massive g
Reddit is going to shit (Score:2)
Re: (Score:1)
Is parent an AI bot posting to slashdot using natural language in which to promote a product? Well played, sir.
Re: (Score:2)
AI is getting help from me to poison Reddit (Score:2)
Since Reddit announced it was going to sell "its" data (i.e. the data it was given for free by all the redditors), I've been busy deleting most of my posting history, and pumping old posts I left there full of technical idiocies and fake facts,
Why you ask?
Because 1/ it hurts Reddit and 2/ it trains AI on bad data.
I have no intention of helping business models based on mooching off people's contributions.
We're Doomed (Score:2)
We're all doomed to drown in a bottomless pit of AI-generated horseshit, brought about by AIs endlessly ingesting and regurgitating other AI-generated horseshit.
Reality will become nebulous, and the only thing you'll be able to trust is whatever was committed to paper before ~1990 or so. After that, everything else and all other 'records', will be suspect.
Maybe you want a formula for some chemical process, will you really be able to trust the information you get online, even if you check multiple sources? W