How to hide AI-generated posts on LinkedIn.

LinkedIn has no setting for "show me fewer posts written by a model". What it has is a handful of per-person controls, a feed you can sort by time, and a growing share of posts that all sound the same. This guide covers what the built-in controls actually do, how to recognise generated text by eye, and how to filter it automatically so you stop doing that by hand.

1. Use LinkedIn's own controls (they work per person, not per post type)

Everything LinkedIn offers is aimed at who you see, not what kind of writing you see. Still worth setting up:

• "I don't want to see this." Open the three-dot menu on any post and pick it. LinkedIn drops that post and treats the choice as a signal for similar ones. It is one post at a time, so it helps most against repeat offenders.
• Unfollow without disconnecting. From the same menu, or from the person's profile, choose Unfollow. You stay connected, they are not notified, and their posts stop appearing. Use it on connections whose feed presence is all recycled quotes.
• Mute. For people and Pages you do not follow but keep seeing because a connection reacted to them. Mute removes that indirect route.
• Prune what you follow. Under My Network, open Following & followers and click Following next to anyone you would rather not hear from. Ten minutes here does more than a month of scrolling past.
• Sort by Recent. The feed defaults to Top, which is where suggested posts and engagement-optimised content concentrate. Switching to Recent shows what your network actually posted, in order.

The limits: none of these look at the text. A connection can post one useful thing a week and six generated ones, and LinkedIn's controls make you choose between all or nothing. Suggested posts from people you have never heard of keep arriving regardless.

2. Learn the tells

Generated LinkedIn posts share a small set of habits. None is proof on its own, but three or four together rarely come from a person typing into the composer:

• The contrast hook. "It's not about X. It's about Y." Often stacked three deep.
• The corporate opener. "In today's fast-paced world…", "In the ever-evolving landscape of…".
• Em-dash stacks. Several long dashes in a short post. Humans on LinkedIn mostly type hyphens.
• Curly quotes. LinkedIn's composer produces straight quotes. Curly ones were pasted in from somewhere else.
• Staccato paragraphs. Every paragraph is one sentence. Every sentence is a line. Nothing connects.
• Emoji bullets. A rocket, a light bulb and a target introducing three "key takeaways".
• The fake arc. "I got rejected 47 times. Then I changed one thing." A setback, a turning point, a lesson, in under 60 words.
• The engagement closer. "Agree?", "Thoughts?", "Repost if this resonated." The post exists to be interacted with, not read.

Once you see these, you cannot unsee them. The problem is that spotting them still costs you the read. Which is the argument for the next step.

3. Filter automatically with a Chrome extension

NoSlop for LinkedIn is a free Chrome extension that scores every post in your feed and collapses the ones that read like slop into a one-line card with the reasons spelled out. Out of the box it runs eleven pattern rules, the tells above, entirely on your device: no account, no network calls, one permission.

Turn on Slop Score and every post also gets a small "% AI" badge. The percentage comes from Jev, a decision model from TypeSafe AI that answers calibrated yes/no questions on four dimensions: AI-written, engagement bait, fake story, and recruiter or sales spam. You choose which of the four should hide a post. It is opt-in behind a consent card, sends only the post text, and stores nothing.

Not listed yet? Join the waitlist for one email when it is live.

4. Set a threshold you can live with

A filter you keep overriding is a filter you will switch off. NoSlop has one slider with three bands. Lenient hides only posts that score 80% or higher: the obvious ones. Balanced, the default, draws the line at 65%. Strict hides anything at 50% or above and will catch some heavily polished human writing too. Start at Balanced for a week, look at what it hid, and move the slider once.

5. Keep the human ones

Every hidden post stays one click away behind "Show anyway", and the card tells you why it was hidden. Scores are probabilities, not verdicts: a terse engineer who writes in short lines can read as generated, and a well-edited human essay can score in the forties. Treat the badge as a reading aid, not a judgement of the author.

Frequently asked

Does LinkedIn have a built-in AI filter? No. As of September 2026 the feed controls are per person (unfollow, mute, "I don't want to see this") plus the Top/Recent sort. Nothing looks at the writing itself.

Does hiding posts affect the author? No. Everything above changes what your browser shows you. Nothing is sent to LinkedIn or to the person who posted.

Can I do this on mobile? LinkedIn's own controls work in the app. Browser extensions do not run in the LinkedIn mobile app, so the automatic filter is desktop Chrome for now.

Questions or a post the filter got wrong: hello@noslop.lol. What the extension does with data is on the privacy page.