X Live-Tweets Its Fight Against Chatbot Spam In Real-Time

Nikita Bier, caput of merchandise astatine X (formerly known arsenic Twitter) posted a bid of bonzer tweets astir spam connected Twitter, explaining the information of immoderate of the spammers, they described types of spam and mentioned that immoderate spammers were utilizing Grok to auto-post spam responses astatine scale.

Nikita Bier spent 24 hours tweeting anti-spam actions successful existent time, revealing specifications astir really how automated and fast-paced AI-assisted level spamming has become.

42,000 Accounts In One Sweep

Bier started their bid of tweets describing the scope of the chatbot problem, saying that providing authenticity connected X is simply a halfway worth that’s cardinal to the company’s halfway values.

Bier tweeted:

“We recovered 42,000 accounts automating replies utilizing chatbots and person removed them from the platform.

X’s halfway worth is providing an authentic beat connected humanity — and utilizing AI to programmatically prosecute pinch users without a quality successful the loop runs antagonistic to our mission.”

Chatbot Spam Is Motivated By Money

Bier followed up the adjacent time pinch a explanation astir what the chatbot spamming was about. They explained that the spam wasn’t ideologically driven, it was not governmental nor a a state-run power operation. It was purely astir monetization. Apparently the spammers were trying to turn an assemblage successful bid to leverage that for paid promotion deals from AI companies looking to widen their influence.

Companies don’t usually station astir their spam issues and the real-time posting by Bier offered a behind-the-scene look astatine the standard and motivations.

Bier tweeted:

Perhaps you should find a different app pic.twitter.com/z9Qv0vbpZj

— Nikita Bier (@nikitabier) July 24, 2026

And besides posted an explanation:

“For transparency, the bulk of them were spamming thought-leadership slop astir artificial intelligence — to turn accounts and person paid promotion offers from AI tech companies.

99.99% of spam connected X is economically-motivated.

Just plain aged grifters.”

The Status Update: A Fix successful Hours, Not Days

After a time Bier posted a speedy update astir a timeline for erstwhile to spot the X provender improve, again showing really these kinds of anti-spam actions activity successful real-time.

Bier tweeted:

“This spam onslaught was mitigated tonight. Your provender will amended successful the adjacent 6-12 hours.”

X Fights Back Against AI Chatbot Spammers

Bier’s tweets served arsenic an unfastened connection to spammers to show really superior they are astir fighting inauthentic behaviour connected the platform, calling the spammers criminals.

Bier tweeted:

“We will cleanable this place. I don’t attraction really galore enemies I create. X will not beryllium manipulated by criminals.”

The Spammer Who’s Pivoted 40 Times

The astir eye-opening portion of the thread wasn’t a short speech betwixt Bier and 1 of the X members. Bier described 1 spam usability that was truthful persistent and adaptive (40 method changes successful six months), that the spammer behaves little for illustration a bot workplace and much for illustration personification who’s sitting correct adjacent to them search each their responses and quickly coming up pinch a countermeasure.

A reply from an X personnel called attraction to the truth that immoderate of these spammers person started utilizing Grok, X’s ain AI product, to make their replies. Bier confirmed that the Grok-based method had already been caught and blocked. In their response, Bier revealed that X’s consequence clip of 12-18 hr turnarounds was an betterment complete really agelong the aforesaid problem utilized to linger nether the aged Twitter.

Bier tweeted:

“There are a fewer spammers connected X that person been pivoting their strategy for the past 6 months.

One of the them (“This feline is simply a awesome trader ⬇️”) has pivoted a full of 40 times aft each method has been blocked.

Some of the techniques are truthful imaginative and accelerated that it feels for illustration they’re sitting correct adjacent to us.

At this point, we mightiness arsenic good prosecute them because they are conscionable arsenic acquainted pinch the X codebase arsenic us.”

@CryptoParadyme responded:

“please do not promote them

a batch of them person been utilizing immoderate benignant of grok reply recently.”

Bier tweeted:

“We blocked the Grok 1 yesterday.

Our squad is opinionated by waiting for their adjacent move. We are 10x much proactive than before.

This would fester for months astatine Twitter but our turnaround clip now is 12-18 hours.”

False Positive Reported And Dealt With

Another absorbing consequence of this thread astir spam actions successful real-time is that 1 personification posted astir their false-positive acquisition pinch the spam actions. A mendacious affirmative is erstwhile a instrumentality makes a mistake, labeling thing arsenic spam erstwhile successful truth it’s not. In existent life, these kinds of algorithms trust connected a multitude of signals successful bid to pinpoint spam pinch a precocious level of accuracy but mendacious positives tin still happen, ideally a debased percent of time.

@the_defi_dad tweeted:

“I gotta beryllium honest. I was amazed to get a announcement for my relationship being spam. I clicked the petition reappraisal fastener and astir 12 hours later I was reinstated. I did kick astir writer app being a scam and instantly sewage a notice. Not judge the nexus location but gladsome to spot the algorithm aliases immoderate decides spam aliases nary spam made the correct call.”

Bier’s posts received a affirmative consequence from X users, immoderate of whom expressed that immoderate of these engagement chatbots must beryllium generating engagement numbers but that their inauthentic quality makes them a scourge.

@RandomPerson242 tweeted:

“On 1 hand, group intelligibly for illustration this rubbish somehow. On the different hand, I work together you can’t fto that beryllium the site’s content, moreover if group travel it. It can’t beryllium a title to the bottom.”

Social Media Is Best When Authentic

Many were happy to spot that X was fighting backmost against inauthentic AI chatbot spam. It ruins the acquisition because group travel to societal media platforms for illustration X to publication and stock quality experiences.

Featured Image by Shutterstock/Naumova Marina