Linkedin no longer wants purely AI generated Slop content. This comes as LinkedIn recently introduced an anti-slop button and soon will let users know when others think they’re starting to sound like bots
Today, nearly every social platform has taken steps to constrain the rapid growth of AI-generated content, primarily through automated labeling tools with uneven track records. The issue is more critical for LinkedIn than social-media firms that focus on entertainment and familiar connections, because the company positions itself as an invaluable source of information and career opportunities for its 1 billion-plus members.
People who post AI-generated material on LinkedIn often hope the tech will make them better communicators, but can wind up sounding generic or inauthentic instead, executives and observers say.
”Everybody uses AI to some degree,” said Brendan Gahan, co-founder and chief executive officer of Creator Authority, a firm that helps businesses and influencers promote themselves on LinkedIn. The firm advises against relying too heavily on AI because success on LinkedIn ultimately comes down to credibility, Gahan said.
AI is hardly the first content challenge for LinkedIn, which has previously confronted oversharing among some users, a wave of scam listings and “engagement pods,” or networks of influencers that artificially boost each other’s reach. “We have been working on this domain of low-quality content for a long time,” said Oscar Rodriguez, LinkedIn’s vice president of product.
According to research, 41% percent of long-form posts and 30% of all comments posted publicly on LinkedIn from April to June were entirely AI-generated, according to an analysis of nearly 57,000 items from AI detection startup Pangram Labs.
Another AI-detection startup, Originality.ai, examined 5,000 public LinkedIn posts in July for signs they used AI more than a “moderate” amount. It concluded that 81% likely met that threshold. Just under half of 5,500 public Reddit posts examined from July fell into the same category.
LinkedIn disputed the third-party vendors’ numbers, but declined to provide its own comparable data.
“Their research appears to treat any content that AI touches as AI slop, which is not how we look at it,” said a LinkedIn spokeswoman. “… AI can be a great support tool in helping people in articulating ideas, refining language or making language more concise.”
The most common red flag is an articulate comment that merely summarizes the post it’s addressing, said Dan Roth, LinkedIn’s editor in chief and vice president of content. “A lot of this content ends up in the uncanny valley of ‘Sounds like humans, but isn’t quite human, and this is giving me a bad feeling’,” he said.
In addition to adding a slop button, LinkedIn last month said it was phasing out an “enhance your post” feature that encouraged the use of AI tools for proofreading and rewriting. The move stemmed from low use and a desire to avoid contributing to the AI problem, according to Roth.
The company plans to introduce a “nudge” feature alerting people when others flag their posts as AI-like.
LinkedIn says it quashes more than 200,000 AI-generated spam comments every day before they go live. It also correctly detects an estimated 94% of generic, or AI-generated, posts, which it refrains from distributing beyond the author’s immediate network, according to a spokeswoman.
LinkedIn is ultimately less concerned with whether a given post uses AI than whether others think it’s worth reading.
“How can we make sure that your time on LinkedIn is a good investment, that when you open the app, you are getting something that makes you better at your job or the job that you wish you had?” said Roth.
In one sign of progress, year-over-year growth in the share of LinkedIn content that’s likely more than moderately AI-generated was 15.9% in the period from January through July this year, down from 20.3% in 2025 as a whole, according to Originality.ai’s analysis of a sample of posts.
Precisely measuring AI-generated or -assisted content has become nearly impossible, however, as the tools advance and spread, said Jérémy Boissinot, CEO and co-founder of Favikon, a platform that indexes and analyzes creator profiles across networks including LinkedIn.
“Obviously, a big share of LinkedIn’s content is AI-assisted, but nobody can know exactly how much, including LinkedIn, and anyone claiming they do is selling a detection tool,” he said.
