Nikita Bier, head of product at X (previously often known as Twitter) posted a sequence of extraordinary tweets about spam on Twitter, explaining the motivation of a few of the spammers, they described varieties of spam and talked about that some spammers had been utilizing Grok to auto-post spam responses at scale.
Nikita Bier spent 24 hours tweeting anti-spam actions in actual time, revealing details about how how automated and fast-paced AI-assisted platform spamming has develop into.
42,000 Accounts In One Sweep
Bier began their sequence of tweets describing the scope of the chatbot drawback, saying that offering authenticity on X is a core worth that’s central to the firm’s core values.
Bier tweeted:
“We discovered 42,000 accounts automating replies utilizing chatbots and have eliminated them from the platform.
X’s core worth is offering an genuine pulse on humanity — and utilizing AI to programmatically have interaction with customers with out a human in the loop runs counter to our mission.”
Chatbot Spam Is Motivated By Cash
Bier adopted up the subsequent day with a clarification about what the chatbot spamming was about. They defined that the spam wasn’t ideologically pushed, it was not political nor a a state-run affect operation. It was purely about monetization. Apparently the spammers had been attempting to develop an viewers so as to leverage that for paid promotion offers from AI firms wanting to lengthen their affect.
Corporations don’t usually put up about their spam points and the real-time posting by Bier supplied a behind-the-scene take a look at the scale and motivations.
Bier tweeted:
Maybe it’s best to discover a totally different app pic.twitter.com/z9Qv0vbpZj
— Nikita Bier (@nikitabier) July 24, 2026
And likewise posted an explanation:
“For transparency, the bulk of them had been spamming thought-leadership slop about synthetic intelligence — to develop accounts and obtain paid promotion gives from AI tech firms.
99.99% of spam on X is economically-motivated.
Simply plain previous grifters.”
The Standing Replace: A Repair in Hours, Not Days
After a day Bier posted a fast replace a couple of timeline for when to see the X feed enhance, once more exhibiting how these sorts of anti-spam actions work in real-time.
Bier tweeted:
“This spam assault was mitigated tonight. Your feed will enhance in the subsequent 6-12 hours.”
X Fights Again In opposition to AI Chatbot Spammers
Bier’s tweets served as an open assertion to spammers to present how critical they are about preventing inauthentic habits on the platform, calling the spammers criminals.
Bier tweeted:
“We are going to clear this place. I don’t care what number of enemies I create. X will not be manipulated by criminals.”
The Spammer Who’s Pivoted 40 Occasions
Probably the most eye-opening a part of the thread wasn’t a brief alternate between Bier and one in all the X members. Bier described one spam operator that was so persistent and adaptive (40 methodology adjustments in six months), that the spammer behaves much less like a bot farm and extra like somebody who’s sitting proper subsequent to them monitoring all their responses and quickly arising with a countermeasure.
A reply from an X member known as consideration to the incontrovertible fact that a few of these spammers have began utilizing Grok, X’s personal AI product, to generate their replies. Bier confirmed that the Grok-based methodology had already been caught and blocked. Of their response, Bier revealed that X’s response time of 12-18 hour turnarounds was an enchancment over how lengthy the identical drawback used to linger below the previous Twitter.
Bier tweeted:
“There are a number of spammers on X which have been pivoting their technique for the final 6 months.
Considered one of the them (“This man is a fantastic dealer ⬇️”) has pivoted a complete of 40 occasions after every methodology has been blocked.
A few of the methods are so artistic and quick that it seems like they’re sitting proper subsequent to us.
At this level, we would as effectively rent them as a result of they are simply as acquainted with the X codebase as us.”
@CryptoParadyme responded:
“please do not encourage them
a whole lot of them have been utilizing some type of grok reply just lately.”
Bier tweeted:
“We blocked the Grok one yesterday.
Our crew is standing by ready for his or her subsequent transfer. We are 10x extra proactive than before.
This would fester for months at Twitter however our turnaround time now is 12-18 hours.”
False Constructive Reported And Dealt With
One other fascinating results of this thread about spam actions in real-time is that one individual posted about their false-positive expertise with the spam actions. A false constructive is when a machine makes a mistake, labeling one thing as spam when in reality it’s not. In actual life, these sorts of algorithms rely on a mess of indicators so as to pinpoint spam with a excessive degree of accuracy however false positives can nonetheless occur, ideally a low proportion of time.
@the_defi_dad tweeted:
“I gotta be trustworthy. I used to be stunned to get a discover for my account being spam. I clicked the request overview button and about 12 hours later I used to be reinstated. I did complain about grass app being a rip-off and instantly bought a discover. Undecided the hyperlink there however glad to see the algorithm or no matter decides spam or no spam made the proper name.”
Bier’s posts obtained a constructive response from X customers, a few of whom expressed that a few of these engagement chatbots have to be producing engagement numbers however that their inauthentic nature makes them a scourge.
@RandomPerson242 tweeted:
“On one hand, folks clearly like this garbage by some means. On the different hand, I agree you may’t let that be the web site’s content material, even when folks comply with it. It could possibly’t be a race to the backside.”
Social Media Is Finest When Genuine
Many had been comfortable to see that X was preventing again in opposition to inauthentic AI chatbot spam. It ruins the expertise as a result of folks come to social media platforms like X to learn and share human experiences.
Featured Picture by Shutterstock/Naumova Marina
Disclaimer: This article is sourced from external platforms. OverBeta has not independently verified the information. Readers are advised to verify details before relying on them.
