- Is it still worth publishing AI content on LinkedIn?
- What has really happened on LinkedIn in recent months?
- What changes in practice for those who manage company content on LinkedIn?
- Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
- Which tools and approaches can you use to produce safe LinkedIn content?
- What are the limits of the report button you need to know about?
- Why everyone is talking about it now: is LinkedIn really relevant for Italian companies?
- How do you get started: the steps to adapt your LinkedIn strategy
- What mistakes are most common during this transition?
- How do I know if my LinkedIn strategy is working after these changes?
- So, does the AI slop button really change anything for your company?
- How we see it at SHM
- Frequently Asked Questions
- On July 30, 2026, LinkedIn launched the 'Seems like AI slop' button: in three weeks it received over a million clicks.
- Posts classified as AI slop lose 40% of their views and are limited beyond the author's direct network.
- 41% of longform posts on LinkedIn were entirely AI-generated before the intervention: the platform responded with new automated classifiers and the removal of the 'Enhance your post' feature.
- For Italian companies, the priority is an audit of LinkedIn content and a human editorial review before publishing.
Updated October 29, 2026
One million people pressed LinkedIn's "Seems like AI slop" button in less than three weeks. If your company publishes content on the platform, this number concerns you directly.
Is it still worth publishing AI content on LinkedIn?
It depends on how you produce it. Content generated automatically without human review now risks a real penalty on organic reach, meaning the visibility you get without paying.
LinkedIn has changed the rules of the game. Since July 30, 2026, every user can report a post as AI slop, a term for content generated automatically by artificial intelligence with no authentic value. The result is immediate: posts classified this way lose 40% of their views, according to Chief Product Officer Hari Srinivasan, speaking to The Register.
For a marketing manager or business owner, this means one thing: publishing volume matters less than it used to. Editorial quality has become a measurable competitive factor.
| Your situation | Current risk | why |
|---|---|---|
| Content written by people, with a recognizable voice | Yes Low | Hard to classify as AI, authentic engagement |
| AI content with human editorial review | Yes Low-medium | An original voice reduces the risk of being reported |
| AI content without review, standardized structure | Not for now High | Automated classifiers at 94% accuracy in internal tests |
| Automated comments at scale | Not for now Very high | LinkedIn blocks hundreds of thousands of attempts every day |
What has really happened on LinkedIn in recent months?
LinkedIn found that almost half of its long-form content was AI-generated, and responded with three simultaneous measures: a reporting button, new automated classifiers and the removal of a tool that made mass production easier.
It all started with a study by Pangram Labs, published in July 2026, which analyzed over one million posts across five social platforms. The result for LinkedIn was alarming: 41% of long-form posts, those over 250 words, were found to be entirely AI-generated. The story was reported by 404 Media and sparked wide debate in the industry.
LinkedIn isn't the only platform affected, but it's the most exposed. According to the same Pangram analysis, LinkedIn accounts for about 62% of all AI content detected across the platforms scanned, almost double any other, as reported by Vynce Digital.
LinkedIn's response came on July 30, 2026, with three parallel measures. First: the Seems like AI slop button in the three-dot menu of every post. Second: new automated classifiers that identify AI-generated content with a claimed accuracy of 94% in internal tests. Third: the removal of the Enhance your post feature, which rewrote posts with AI without preserving the author's voice, replaced by a proofreading tool that corrects the text without altering its tone, as reported by Fortune.
What changes in practice for those who manage company content on LinkedIn?
Content produced automatically without human oversight loses visibility and risks being reported. The distinction between automating the process and automating editorial thinking becomes the most important dividing line.
The most significant change isn't the button itself, but the combination of human reporting and automated classification. Posts identified as AI slop are limited in distribution beyond the author's direct network: in practice, they stop circulating outside your circle of direct connections. For a company that uses LinkedIn to generate business leads, this is real damage to organic reach.
There's a second front too: comments. According to a Pangram analysis of 57,000 public posts (April–June 2026), 30% of LinkedIn comments were entirely AI-generated. LinkedIn responded with a new comment ranking system based on professional relevance, as Social Day reports. Generic automated comments, like "Great article, very interesting!", lose visibility.
- 1 · AI / toolResearch and structureAI gathers data, suggests angles, and produces an organized draft.
- 2 · Human editorPoint of view and voiceThe professional adds an original perspective, real examples, and a recognizable tone.
- 3 · Human editorPre-publication reviewQuality check: does the content sound like a real person? Does it add value?
- 4 · PlatformAutomatic classificationLinkedIn analyzes the post with the new classifiers. Stated accuracy: 94% on generic content.
- 5 · CommunityCollective reportingUsers can report the post as AI slop through the three-dot menu.
Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
The impact varies widely depending on publishing volume and content type. If you post rarely but with an authentic voice, you're already safe. If you post a lot on autopilot, you need to act now.
These scenarios are illustrative: the numbers show the order of magnitude of the risk and are not real client data.
| Company type | Typical situation | Current risk | Priority action |
|---|---|---|---|
| E-commerce (fashion or furniture) | 20-30 AI-generated posts/month, little review | High: standardized content, easy to classify | Immediate audit + editorial review on every post |
| Professional firm (accountant, lawyer) | 4-8 posts/month, written by the professional or with AI + review | Low: recognizable personal voice | Keep up the oversight, add real examples |
| B2B manufacturing company | Technical product posts, often delegated to an agency | Medium: depends on the agency's workflow | Check the agency's process, ask for documented human review |
Which tools and approaches can you use to produce safe LinkedIn content?
It's not about choosing between AI and human writing. It's about building a process where AI supports and humans decide. Different tools serve different stages.
| Approach | What AI does | What humans do | AI slop risk |
|---|---|---|---|
| AI as a full replacement | Writes, publishes, comments automatically | None | Not for now Very high |
| AI as a draft, light human review | Produces the structure and base text | Edits and approves in 10 minutes | Yes Medium-low |
| AI for research, human writing | Gathers data, suggests angles | Writes the final text | Yes Low |
| Entirely human writing | No role | Writes everything | Yes Minimal |
LinkedIn's new proofreading tool, which replaces the removed Enhance your post feature, is an example of AI used the right way: it fixes errors without changing your voice. This is exactly the line companies need to learn to manage internally.
If you run LinkedIn campaigns in a structured way, the LinkedIn checklist for B2B campaigns offers an operational reference updated for 2026.
What are the limits of the report button you need to know about?
The collective reporting system has three weak points: the subjectivity of judgment, the risk of competitive abuse, and the algorithmic variable that companies don't control.
The perception of AI slop is subjective. Content written in a very formal style or with a very linear structure can be reported even if it's entirely human. LinkedIn hasn't yet clarified publicly how it will handle systematic abuse, for example coordinated reports from competitors.
The most relevant point from a strategic standpoint is something else: generative AI doesn't necessarily produce low-quality content. It produces content with no original perspective when used without a solid editorial framework. The problem isn't the tool, it's the method.
The updated automatic classification introduces an algorithmic variable that companies can't control directly. That makes it even more important to invest in a content strategy that produces recognizably human, contextually relevant text. To see how LinkedIn is evolving on internal tools too, read our analysis LinkedIn vs. AI slop: reporting and proofreading.
Why everyone is talking about it now: is LinkedIn really relevant for Italian companies?
With 28.2 million registered users in Italy as of September 2026, or 47.3% of the population, LinkedIn is the go-to professional platform for Italian B2B. Ignoring it isn't an option.
The growth of Italian LinkedIn users is steady. According to NapoleonCat data, in 15 months (June 2025 to September 2026) the Italian user base grew from 24 to 28.2 million: an increase of 17.5%. That means your professional audience is more and more present on the platform, and more and more exposed to algorithm changes.
A growing platform that changes its editorial rules is a signal to take seriously. If you adapt your strategy now, you get ahead of those who wait for the problem to show up in performance data. The topic ties into the broader issue of AI content across all platforms: for a comparison, the article LinkedIn and AI-generated content: 41% of posts are artificial gives the full context of the Pangram study.
How do you get started: the steps to adapt your LinkedIn strategy
The starting point is an audit of the content you've published over the past six months. From there, you build an editorial process that clearly separates the role of AI from the role of humans.
- 1 · Marketing managerAudit your recent contentReview your posts from the last 6 months: how many were produced with fully automated workflows? How does their engagement rate compare with content written by people?
- 2 · Editorial teamDefine internal guidelinesSeparate AI as a support tool (research, drafts, headlines) from AI as a replacement for editorial thinking. Put in writing what AI can and can't do.
- 3 · Editor / content leadIntroduce pre-publication reviewEven 10 minutes of human review before every post makes a difference in perceived quality. The voice has to sound like a real person.
- 4 · Marketing managerRethink your content KPIsThe number of posts published matters less. Track qualitative engagement rate, follower growth and reach beyond your direct network instead.
- 5 · TeamFocus on formats that are hard to replicate with AIPersonal posts from founders and managers, analyses based on proprietary data, content that documents real internal processes: these formats need a human contribution AI can't replace.
What mistakes are most common during this transition?
The most common mistake is thinking it's enough to post less often. The problem isn't how much you publish, but how you produce each individual piece of content.
Here are the three mistakes we see most often among Italian companies that manage LinkedIn in a structured way:
- Confusing form with substance. Adding emojis, breaking up paragraphs or using an informal tone doesn't make AI content any less recognizable. Classifiers look at logical structure and the presence of an original perspective, not formatting.
- Handing everything to the agency without checking the process. If your agency produces LinkedIn content automatically with no human review, the risk is yours. Ask explicitly how every post is produced.
- Ignoring comments. With 30% of LinkedIn comments classified as AI-generated, even automated replies to company posts can damage how your brand is perceived. Generic comments with no context are now downranked by the algorithm.
How do I know if my LinkedIn strategy is working after these changes?
The metrics to track are changing: publishing volume gives way to qualitative engagement and reach beyond your direct network.
| Indicator | What it measures | Positive signal | Red flag |
|---|---|---|---|
| Reach beyond your direct network | Distribution outside your direct connections | Stable or growing | Sudden drop after a post |
| Engagement rate (likes + comments / impressions) | Perceived content quality | Above industry average | Below 1% at high volumes |
| Quality of comments received | Community relevance | Specific, contextual comments | Generic or automated comments |
| Reports received | AI perception of the content | Zero reports | Repeated reports on the same type of post |
| Qualified follower growth | Editorial appeal | Steady growth within your target | Stagnation or decline |
So, does the AI slop button really change anything for your company?
Worth it if you already have a LinkedIn content strategy with human editorial review: the change puts you ahead of competitors who publish on autopilot.
Somewhere in between: you use AI to produce content but have a review step. Revisit your process, strengthen your editorial voice and monitor your reach metrics.
Not yet if you publish fully automated content with no oversight: the risk of being penalized is real and immediate. Start with the audit.
How we see it at SHM
For us, AI is a tool that gives time back, not a substitute for editorial ingenuity. What's happening on LinkedIn confirms it: platforms are learning to tell content produced with a method from content produced in volume. And the difference shows in reach.
In the content strategy projects we run, human editorial review has never been optional. Not because AI isn't useful (it absolutely is), but because point of view, angle and contextual relevance are human skills that no classifier can generate. That's true for LinkedIn and for any other channel.
If you want to understand how to adapt your B2B content strategy to these changes, you can start with a review of your LinkedIn editorial plan or an analysis of your active campaigns. Our LinkedIn campaigns and strategic AI integration services are designed for exactly this: using technology where it helps, and putting human intelligence where it counts. For a full picture of the opportunities and risks in the Marketing and performance ADV cluster, you'll find all our updated analyses in the dedicated section.
It's also worth looking beyond LinkedIn: low-quality AI content affects every channel, and the strategic answer is always the same: build a recognizable editorial voice that no algorithm can mistake for automated. If you also want to explore paid campaigns on other channels, our analysis of TikTok premium formats offers a useful comparison of how platforms are evolving their editorial policies.
Frequently Asked Questions
Yes, the risk exists. LinkedIn hasn't yet publicly clarified how it will handle cases of systematic abuse. The best defense is to produce content with a recognizable editorial voice and an original perspective: it's structurally harder to classify as AI, even for automated systems.
No, if the text is written by a person and AI only handles spelling or grammar correction. LinkedIn's new proofreading tool is built on exactly this principle: it corrects without altering your voice. The risk arises when AI rewrites or generates the content.
Even ten minutes is enough for a standard post. The goal isn't to rewrite everything, but to check that the content has an original perspective, real examples and a recognizable tone. An internal checklist of 5 questions can make the process systematic without weighing down your workflow.
Yes. According to a Pangram analysis of 57,000 posts, 30% of LinkedIn comments were entirely AI-generated. LinkedIn responded with a new ranking system that downgrades generic comments. Automated comments with no context hurt brand perception and get penalized by the algorithm.
There's no direct notification for the author. The clearest indirect signal is a sudden drop in reach beyond your direct network after publishing a post. Monitoring this metric, along with your engagement rate, is the most practical way to tell whether the algorithm is limiting distribution.
Content that's already published can still be reported by users through the button. The new automated classifiers scan the feed continuously, not just new posts. If you have older, automatically produced content with high visibility, it's worth considering whether to archive or update it.
The real problem isn't AI: it's publishing content you wouldn't have written yourself.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
