What is AI slop? How to spot it in your feed

AI slop is low quality content produced in bulk, usually with generative AI, to farm attention rather than to inform. In a social feed it looks like engagement bait, recycled templates, empty motivation, ragebait, and confident advice with nothing behind it. The useful test is not whether a model wrote the post. It is whether the post gives you anything at all.

The word spread fast enough that Merriam-Webster made slop its 2025 word of the year. It defined slop as "digital content of low quality that is produced usually in quantity by means of artificial intelligence." This guide covers what that looks like on LinkedIn, X, and Reddit, how to recognise it quickly, and what you can do to see less of it.

Where the word slop comes from

Slop originally meant soft mud, and later the leftover food fed to pigs. Its modern sense, cheap content made in volume, took hold as generative AI made writing a post nearly free. When a post costs nothing to produce, the only thing worth optimising is how much attention it gets. Slop is what that incentive produces.

That is why slop is not only an AI problem. Ghostwritten "thought leadership", copy and paste growth hacks, and viral templates existed before chat models did. AI simply made them cheaper and faster to produce.

Slop is a behaviour, not an author

A common mistake is to treat AI slop and AI writing as the same thing. They are not.

• A founder who uses a model to tidy up a detailed account of a failed launch has written a good post with help.
• A person who types "Agree? 👇" under a recycled quote by hand has written slop without any help at all.

That is why AI detectors are the wrong tool here. They estimate whether text was machine generated, a question they often answer badly. Even when they are right, they answer the wrong question. What you want to filter is posts that waste your time, whoever wrote them.

The 8 kinds of AI slop in your feed

These are the patterns SlopStamp's default policy looks for. They overlap, and the worst posts manage several at once.

1. Engagement bait

The post asks for likes, comments, reposts, or a one word reply, and gives you nothing for it. "Comment YES and I will send you the guide." "Agree?" "Repost if you have ever felt this." The reply is the point. Read the full breakdown in LinkedIn engagement bait: 8 patterns to spot.

2. Recycled template

The familiar creator format: a shocking one line hook, short stacked paragraphs, a lesson, a moral. "I got fired on a Tuesday. Here is what it taught me about leadership." The story changes and the shape never does.

3. Generic motivation

Inspiration with nothing underneath it. No experience, no evidence, no detail. "Discipline beats motivation." "Your network is your net worth." You have read it before and you will read it again tomorrow.

4. Ragebait

A hot take or sweeping claim about a group, written to start a fight in the comments. "Gen Z does not want to work." "Remote workers are stealing from their employers." The anger is the engagement.

5. Unsupported expertise

Confident authority that you cannot check. "After advising 300 companies, I can tell you the only metric that matters." There are no names, no numbers you could verify, and no reasoning.

6. Disguised self promotion

An ad dressed as advice. It holds back exactly what you would need to act on it ("DM me for the framework") so that the next step is a sales conversation.

7. Low information posts

Vague, obvious, or falsely profound. "Leadership is not a title. It is a choice." It sounds like insight and contains none.

8. Formulaic AI style

Repetitive phrasing and formatting that drown out whatever the point was. Rows of emoji bullets, every sentence its own paragraph, the same "It is not X. It is Y." construction four times in a row. This is the one category where the AI tells actually matter, and even here the problem is the noise, not the tool.

How to spot slop in three seconds

You do not need a detector. Ask three questions as you scroll.

1. Is there a specific detail? A number you could check, a named situation, a concrete step. Slop avoids specifics because specifics can be wrong.
2. What does the post want from me? If the answer is a comment, a repost, or a DM, and it gives you nothing first, it is bait.
3. Could anyone have written it? If you could swap the author's name for any other creator's and nothing would change, it is a template.

Two "no" answers and one "a comment" usually mean slop. If you want a second opinion on a specific post, paste it into Slop Check. It is free, needs no account, and returns a verdict with the reason.

What each platform lets you do about it

Every major platform gives you some control. None of them filters on what a post says.

• LinkedIn offers Not interested on a post, Hide posts from an author, and Unfollow, as described in its help center. There is no keyword filter. The full walkthrough is in How to hide AI posts on LinkedIn.
• X has muted words, which remove posts containing a word, phrase, or hashtag from your home timeline and notifications. It is the most powerful built in tool of the three, but it only matches exact words.
• Reddit lets you choose Show fewer posts like this and mute whole communities, which keeps them out of your Home and Popular feeds.

The shared weakness is that these controls work on authors, communities, or exact words. Slop is a pattern, so it slips past all three. A new account posts the same template, the bait changes from "Agree?" to "Thoughts?", and you are back where you started.

Filtering slop by rule

The fix is a filter that reads each post and judges it against a description you write in plain language. "Hide posts that ask for comments without giving anything" covers every version of the bait, including ones nobody has written yet.

SlopStamp does this in Chrome for LinkedIn, X, and Reddit. You start from the default policy above or write your own, and you can set different rules for each platform. Posts that match are hidden or stamped with the reason. Nothing is removed from the platform, and anything hidden is one click away. The setup guide takes about two minutes.

The short version

AI slop is content made to be engaged with rather than read. Spot it by the missing detail, the ask, and the template. Use each platform's controls on the worst offenders, and filter the rest by rule so new slop does not keep slipping through.