By Hezron Ochiel
LinkedIn once encouraged users to improve their posts with artificial intelligence.
In a sudden twist, it is now removing that feature and introducing new ways for members to report content that “seems like AI slop.”
The concern is also supported by evidence. A 2026 analysis by AI-detection company Pangram Labs found that LinkedIn had the highest share of fully AI-generated content among the major text-based platforms it examined. More than 40 percent of LinkedIn posts longer than 250 words were flagged as fully AI-generated.
That helps explain why LinkedIn now sees low-value, mass-produced content as a serious feed-quality problem. The harder task will be defining AI slop without punishing legitimate creators who use AI responsibly.
LinkedIn is not acting alone. Substack has partnered with AI-detection company Pangram to help readers identify AI-written material, while investors are putting millions of dollars into technologies designed to detect machine-generated content. The debate across publishing platforms is shifting from whether AI content exists to how it should be identified, labelled and managed.
AI can help people organise ideas, improve grammar and communicate more clearly. It can also make it easier to produce repetitive posts, automated comments and polished writing that offers little value.
LinkedIn is now building a system in which automated detection, member reports and creator feedback may influence what receives visibility.
What LinkedIn is changing
In a recent post, LinkedIn Chief Product Officer Hari Srinivasan said fighting AI slop had become a top priority for the platform.
He said people come to LinkedIn to hear from real professionals, learn from genuine expertise and build meaningful relationships. He added that mass-produced posts and automated interactions threaten that experience.
LinkedIn’s response goes beyond the reporting button. The company is introducing classifiers to identify suspected slop, reducing its appearance in recommendations, privately alerting creators when users view their posts as inauthentic, strengthening its automation defences and replacing its AI rewriting tool with proofreading designed to preserve the writer’s voice.
Member reports will also provide signals that LinkedIn can use to refine its classifiers. The button therefore serves as a complaints mechanism and a data-collection tool, connecting member perceptions, automated detection and content distribution.
This means that user judgement may help shape the automated systems deciding which posts receive less distribution.
LinkedIn is also expanding profile and company page verification and giving members greater control over unwanted comments on company pages.
That swing matters.
Proofreading supports the writer. Full rewriting can gradually replace the writer’s expression with language that sounds polished, familiar and interchangeable.
LinkedIn says it blocks hundreds of thousands of automated comment attempts each day and has stopped millions of other automation attempts in recent months.
The problem, then, extends beyond AI-assisted writing. It includes the systems being used to manufacture activity, imitate conversation and manipulate visibility.
Creators already recognise the problem
I have previously written about why some LinkedIn creators may be losing reach and engagement.
Creators often blame the algorithm when their posts perform poorly. LinkedIn’s latest statement suggests that its efforts to reduce low-value and automated content may also be affecting some creators’ visibility.
The platform is responding to patterns that audiences are already learning to recognise: dramatic hooks, artificial vulnerability, recycled lessons, predictable formatting and comments that repeat the original post without adding anything useful.
A post can look polished and still feel empty.
When many people rely on similar prompts and templates, individual voices begin to disappear. Professionals from different fields start sounding as though the same person wrote their posts.
LinkedIn’s announcement makes the connection between visibility and content quality clearer. The platform now treats repetitive, automated and low-value material as a threat to the quality of the feed.
Why LinkedIn’s action makes sense
Professional platforms depend on trust.
People use LinkedIn to follow industry discussions, identify potential employees and business partners, and build professional reputations.
But when automated systems produce large volumes of posts, genuine experience becomes harder to recognise.
A professional sharing an original lesson from years of work may find that insight competing with dozens of generic posts produced within minutes. This means that the feed becomes busier without necessarily becoming more useful.
The problem is especially visible in the comments.
Popular posts often attract streams of responses such as: “Great insight.” “This really resonates.” “Such an important reminder.”
These comments create the impression of conversation without its substance. At scale, they can distort how credible, popular or widely supported a post appears.
Several people responding to Srinivasan’s announcement argued that automated comments may be an even bigger problem than AI-assisted posts.
LinkedIn is therefore right to strengthen enforcement against fake interaction and large-scale automation.
AI use is not the real test
Srinivasan made an important distinction: AI-assisted content is not automatically AI slop.
Many professionals use AI to check grammar, shorten a paragraph, organise notes, test a headline or translate an idea. These uses do not automatically reduce the value or authenticity of the final work.
The better test is whether the content contains experience, evidence, judgement and accountability.
Consider two posts.
The first is written entirely by a person, repeats a familiar motivational message, and offers no new insight.
The second is based on original professional experience, supported by evidence and refined with help from an AI tool.
The first may be human-written and still be poor. The second may be AI-assisted and still provide genuine value.
As one commenter observed, slop remains slop regardless of how it was produced.
LinkedIn will therefore need to judge more than the suspected method of production.
Where the system could go wrong
Allowing members to report suspected AI slop could help LinkedIn understand what users experience as repetitive, misleading or inauthentic.
The reporting tool could also be misused.
“This seems like AI slop” could become another way of saying: “I disagree with this.” “I do not like this person.” “This person is my competitor.”
One commenter compared the risk with coordinated mass reporting in online gaming, where groups can use reporting tools to target particular users. Another asked what protection creators would have if competitors or organised groups deliberately reported their posts.
There is also a risk that polished or formal writing will be treated as evidence of AI use.
Researchers, technical experts and professionals writing in a second language may naturally communicate this way or use grammar tools to improve clarity.
None of this proves that the person handed over their thinking to AI.
LinkedIn will need to detect unusual reporting patterns, limit the influence of bad-faith reports and prevent coordinated groups from suppressing legitimate opinions.
LinkedIn is correcting one of its own experiments
There is an uncomfortable tension in LinkedIn’s announcement. The platform is now correcting one of its own AI experiments.
LinkedIn’s “enhance your post” feature was designed to help members improve their writing. Yet rewriting tools can also standardise expression, especially when large numbers of users receive similar suggestions.
The result may be technically polished content that gradually loses personal voice and variety.
Replacing full rewriting with proofreading is therefore a sensible correction. It also shows that platforms can contribute to the behaviour they later have to regulate.
The definition of AI slop remains unclear
LinkedIn’s greatest challenge is fairly defining poor content.
What exactly makes a post AI slop?
Is it the amount of AI used?
Is it repetition, weak evidence or lack of originality?
Is it automation, factual weakness or deliberate manipulation?
Could it simply mean content that readers find boring?
That common understanding is very important because many users recognise the problem. However, it is unstable because different people use it to describe different things.
LinkedIn would be on firmer ground if it focused on clearer signals such as automation, copied or fabricated content, impersonation, coordinated manipulation and repeated posts that offer little original value.
These behaviours can be assessed more fairly than whether a post merely “sounds like AI.”
A working definition: AI slop is content produced or distributed at scale with little original experience, evidence, judgement or accountability, primarily to create the appearance of useful participation.
The comparison table
|
AI-assisted professional content |
AI slop |
|
Begins with real knowledge or experience |
Begins with a generic prompt |
|
Uses evidence and specific examples |
Relies on broad claims |
|
Reflects the writer’s judgement |
Repeats familiar conclusions |
|
Is reviewed and verified by a person |
Is published with little scrutiny |
|
Uses AI to improve expression |
Uses AI to replace substance |
|
Accepts responsibility for every claim |
Hides behind automated production |
The Human Value Test
The distinction becomes clearer when content is assessed by its value rather than by whether AI was involved.
Creators and communication teams can apply four questions before publishing. Together, they form the Human Value Test.
One approach is the Human Value Test.
1. Experience
Does the post contain something the writer has observed, done, studied or learned?
AI can generate a general lesson. Professional experience gives that lesson context and meaning.
2. Evidence
Does the post include a specific example, result, fact, source or real situation?
Generic claims are easy to produce. Evidence gives the content weight.
3. Judgement
Does the writer offer a clear interpretation?
Professional value often comes from explaining what an event means, why something worked or what should happen next.
4. Accountability
Is the writer prepared to stand behind every claim?
Using AI does not remove responsibility. The person publishing the content remains accountable for its accuracy, fairness and impact.
A post that fails all four tests may sound impressive and still offer very little value.
What this means for communication teams
The debate extends beyond individual creators.
LinkedIn’s changes mean that this pressure may now carry a cost to visibility and reputation when production volume exceeds the organisation’s supply of original knowledge.
Volume should not become the objective.
Communication teams should begin with the knowledge held by employees, leaders and subject-matter experts. AI can then help organise that knowledge, simplify the language or improve its presentation.
Consider a hospital leader preparing a post about patient care. AI may produce a polished message. Unless the post includes an actual experience, decision or lesson from the institution, it could have been published by almost any organisation.
The institution must still provide the facts, examples, context and professional judgement.
Without those elements, AI simply makes weak content easier to produce at scale. The result may be grammatically correct and offer little reputational value.
What creators should do now
LinkedIn’s action should not create panic among responsible creators.
The lesson is to keep AI from replacing the substance of the work.
Begin with something real: an experience, observation, result or lesson.
Use AI to structure the material, test clarity or improve grammar. Verify the information, remove generic language, and ensure the finished post reflects your actual thinking.
A useful final question is: Could this post be published under almost anyone’s name?
When the answer is yes, the post probably needs more evidence, context or personal judgement.
Strong professional content carries signs of ownership. It tells readers what the writer knows, how they know it, why it matters and where they stand.
This is a content-governance experiment
LinkedIn is not merely testing a new reporting feature. It is conducting a content-governance experiment in which member perceptions may help train automated systems that influence visibility.
That makes the definition of slop critically important. The platform should avoid judging content merely because it sounds polished, structured or AI-assisted. It should focus on clearer signals such as mass automation, fabrication, repetition, coordinated manipulation and the absence of original value.
Success will depend on whether LinkedIn can reduce manufactured activity without suppressing genuine professional voices.
The decisive question should not be, “Was AI involved?” It should be, “What human value remains?”
Hezron Ochiel is an award-winning strategic communications and public relations professional with over 15 years of experience in media, digital communication, and reputation strategy. He serves as the Deputy Corporate Communications Manager at the government-owned Kenya Medical Training College (KMTC) and is the founder of Hezron Insights, where he writes about AI visibility, Digital PR, SEO, GEO, and digital authority. His work has appeared on platforms including Reuters, The New Humanitarian, and The Standard.