Bots took over the internet. Now, tech firms are trying to woo back humans
Has the internet reached peak slop?
On LinkedIn, short, choppy sentences are ubiquitous. Cryptocurrency enthusiasts and serial entrepreneurs leverage artificial intelligence-optimized LinkedIn posts — rife with em dashes — to promote their personal brands. Instagram, once known for highly manicured depictions of the human experience, has become an endless scroll for engagement farming and branded content. On X, the social media site owned by the globe’s inaugural trillionaire, chatbots debate with each other over the latest culture war feud dividing humans.
Now, people are pulling back. Can tech companies make their platforms social — once again — for users with heartbeats? After a full-throttled embrace of AI, the backlash is now well underway.
Cue the latest Silicon Valley buzzword: authenticity.
On Friday, X announced the company is once again in the throes of a bot purge. This time, the effort is part of a broader battle for legitimacy among humans. Nikita Bier, the company’s head of product, announced late last week that the company had eliminated some 42,000 accounts known for “automating replies using chatbots.”
“X’s core value is providing an authentic pulse on humanity,” Bier wrote. “And using AI to programmatically engage with users without a human in the loop runs counter to our mission.”
Can X purge AI-generated spam from its platform?
It’s unclear how the company determined the accounts were run by chatbots, or whether it’ll have a measurable improvement on a platform known for spam. But the announcement is part of a broader trend. After tech companies went all-in on AI, they’re increasingly implementing policies that go the other direction for users.
For years, content moderation has centered on countering disinformation and preventing harm to users, such as cyberbullying. Now, companies are shifting their focus from moderating what people say online to policing who — or what — is saying it.
Bier didn’t respond to an interview request from Straight Arrow. But in response to an X account that may or may not be a bot, Bier — who one user lauded as “the Marie Kondo of X” — said he had “personally infiltrated every Telegram group” to track down inauthentic users.
“For transparency, the bulk of them were spamming thought-leadership slop about artificial intelligence — to grow accounts and receive paid promotion offers from AI tech companies,” Bier said. “99.99% of spam on X is economically-motivated. Just plain old grifters.”
How much activity is from AI?
It’s difficult to nail down a precise statistic about the number of bots actively flooding X with AI-generated slop. The social media platform, formerly known as Twitter, once claimed less than 5% of accounts were fake. Other analyses more recently placed the number closer to 15%. A former cyber-operations officer for the Central Intelligence Agency predicted in 2022 that the proliferation of fake users is much, much higher — representing as many as 80% of the website’s accounts.
Either way, X’s purge of 42,000 chatbot accounts may not have a marked benefit for the site’s users, said Filippo Menczer, an informatics and computer science professor at Indiana University Bloomington. Menczer’s research has documented widespread bot activity on X, formerly known as Twitter, where chatbots produce human-like content to push spam, scams and coordinated amplification.
“I would speculate that banning 42,000 accounts is a tiny drop in the ocean when you consider the transparency reports that Twitter used to publish” before Elon Musk acquired the company in 2022, Menczer wrote in an email to Straight Arrow. “Back then, they used to suspend more than that many accounts **per week**, on average. And that was before the explosion of generative AI!”
X is far from the only social media platform with an AI slop problem. In fact, others have it far worse, according to a report released this month by AI detection company Pangram, with career-focused social media website LinkedIn being crowned “the most AI-saturated platform” on the web.
On X, a quarter of posts were written with the help of AI, according to the research, and another quarter were “fully AI-generated.” On LinkedIn, more than 40% of posts were entirely AI-generated, according to the Pangram analysis.
Substack joins AI crackdown
At Substack, the independent newsletter platform where a fifth of posts have been flagged by Pangram as being either AI-generated or AI-assisted, executives used the study’s findings to differentiate their brand. Last week, Substack announced a partnership with Pangram that enables an easy way for users to determine whether a post was AI-generated.
But as Straight Arrow reported, some Substack users have called into question Pangram’s accuracy. Others have argued that the substance of an article, rather than the tools used in its production, is most important.
That hasn’t stopped Substack executives from calling out its users who write their newsletters with help from AI — and from taking on LinkedIn as its punching bag.
“We’re sick of slop and we don’t want substack to turn into LinkedIn,” Chris Best, Substack’s co-founder and CEO, wrote on X last week.
Substack didn’t respond to interview requests from Straight Arrow.
Other companies have also rolled out efforts to deprioritize AI-generated content. While TikTok offers tools that allow users to easily create posts with AI assistance, creators are required to add labels to AI-generated content. Since 2024, AI-generated posts on Meta-owned platforms Facebook, Instagram and Threads have been met with an “AI info” disclosure label.
In June, Meta launched an AI-driven “Creator Assistant” dashboard on Facebook that provides tailored feedback to boost the performance of posts. Earlier this month, the company rolled out a feature on Instagram that allowed users to “remix” other people’s images before backtracking three days later after public outcry over the tool’s privacy implications.
At LinkedIn, an AI-writing tool is integrated directly into the platform, allowing users to refine posts with the assistance of a chatbot. Still, a LinkedIn spokesperson told Straight Arrow the social network’s users turn to the platform “to hear from real people and their unique insights and experiences.” The spokesperson didn’t directly address the shade being thrown its way by Substack.
“We actively work to reduce low quality, automated or generic content,” the spokesperson said. “While AI can be used to beat the blank page problem, our focus is on surfacing professional conversations that help people advance their careers.”
Why combating slop won’t be easy
Back at X, Bier has spent the last few months taking shots at the platform’s biggest fraudsters.
In May, Bier announced that the company would crack down on X users with large followings who game the platform’s revenue-sharing program by “programmatically reuploading content from smaller accounts” without proper attribution. The X revenue-sharing program allows certain content creators to earn money based on engagement.
Bier singled out one X user in particular, a cryptocurrency consultant with 3.7 million followers, after he reposted a video of an ABC News journalist ducking for cover after gunshots rang out outside the White House. A community note accused the crypto consultant of stealing the video “without providing credit” to the original source.
If X is committed to clearing up the platform, Bier is well aware of the work ahead of him. In a post on the platform on Saturday, he highlighted how multiple spam accounts have “been pivoting their strategy” to evade bans.
“Some of the techniques are so creative and fast that it feels like they’re sitting right next to us,” Bier wrote. “At this point, we might as well hire them because they are just as familiar with the X codebase as us.”
Round out your reading
- What does low water at the Hoover Dam mean for power in the American West?
- Could GLP-1 drugs become the first longevity medications? Scientists think it’s possible.
- ChatGPT told him he wouldn’t face charges. Then came the arrest warrant.
- How one small city built a second act as a bio hub inside empty cigarette factories.
- If permanent Daylight Saving Time is so popular, why is it dead in the Senate?
Ella Rae Greene, Editor In Chief
