What is AI slop? Meaning and examples in writing

What “AI slop” means

digital content of low quality that is produced usually in quantity by means of artificial intelligence

That’s the definition Merriam-Webster gave “slop” when it named it the word of 2025. A month later, the American Dialect Society picked “slop” as its word of the year too, for “low-quality, high-quantity content, most typically produced by generative AI.”

Two things make content slop. It’s made with little care, and it’s pushed at people who didn’t want it. Most of the AI slop people talk about is pictures and video. This article is about the written kind: posts, replies, emails and articles.

Where the word came from

People on forums were calling AI images “slop” soon after the image tools arrived in 2022. The developer Simon Willison helped make it the common word with a post in May 2024:

Not all promotional content is spam, and not all AI-generated content is slop. But if it’s mindlessly generated and thrust upon someone who didn’t ask for it, slop is the perfect term for it.

His point still holds. Writing with help from AI isn’t slop by itself. Slop is what you get when nobody read, chose or cared about the words before they went out.

AI slop examples in writing

In replies and comments

Great insights! Thanks for sharing this valuable perspective.

This is the most common kind, because replies are short and easy to send by the hundred. It praises without saying what was good, and it would fit under any post on any subject. A reply that came from reading the post names something in it: “The second chart surprised me. Is that before or after the price change?”

In posts

I’m thrilled to announce that I’ve embarked on a new chapter. Here’s what I learned: 1. Consistency is key. 2. Embrace failure. 3. Never stop learning.

Big feelings, three lessons anyone could write, nothing that happened to this person. A real version says what happened: “I left my job in May. The first client came from a reply I wrote in March.”

In emails

I hope this email finds you well. I wanted to reach out to explore potential synergies between our teams.

Two sentences in, and the reader still doesn’t know what you want. Say it in the first line.

In reviews and product pages

This product is a game changer that seamlessly elevates your daily routine.

No detail a buyer can check. What does it do, and what changed for you?

In articles

In today’s fast-paced digital landscape, it’s more important than ever to delve into the key factors that drive success.

A paragraph that could open any article on any topic. Cut it and start with the second paragraph.

The platforms are pushing back

In 2026, the big sites started treating slop replies as a problem of their own.

• LinkedIn said in May 2026 that it is going after comments posted at scale with automation and “responses that simply restate the original post without sharing anything new”. LinkedIn’s post
• X said in July 2026 that it had found 42,000 accounts “automating replies using chatbots” and removed them from the platform. Search Engine Journal’s report
• LinkedIn added a “Seems like AI slop” option to posts in July 2026, and said it blocks “hundreds of thousands of automated slop comment attempts every day”. Fortune’s report

On X, a reply that reads like slop can also end up behind Show probable spam, where almost nobody sees it.

How to spot AI slop in writing

Ask these about a post, a reply or an email:

1. Could it sit under anything else? If you could paste it under another post and it would still fit, it’s slop.
2. Does it repeat what was said? Replies that restate the post add nothing.
3. Are the words the usual ones? Words like delve, tapestry and testament are on every list. See AI words and phrases everyone recognizes.
4. Is there an em dash in every other line? It was the best known sign of AI for years. See why AI uses so many em dashes.
5. Is there one specific thing in it? A number, a name, a detail from the picture. Slop has none.

One sign proves nothing. People wrote “pivotal” and used em dashes long before chatbots. Several signs together, and no specifics, is when it’s slop.

How to keep slop out of your own writing

• Start from the thing in front of you: the post, the email, the product. Say something only it has.
• Use the plain word. “Shows” works where underscores doesn’t.
• Cut the opener and the sign off. Start with the point and stop when it’s made.
• If you write with AI, read every line before it goes out, and change what doesn’t sound like you.

Replies are where slop does the most damage, because people read them in a second and decide whether you’re worth a follow. That’s the job Twick does: it drafts replies with 100% no AI slop, about the post you’re reading and in your own voice, and you post the one you like yourself. It works on X (Twitter), Threads and Bluesky.

Questions

Is AI slop only pictures and video?

No. Most talk about AI slop is about images and video, but text is slop too when it’s mass produced and says nothing: replies, comments, posts, emails, reviews and articles.

Who came up with the term “AI slop”?

Nobody owns it. It was in use on forums by 2022, and the developer Simon Willison’s post in May 2024 helped make it the word everyone uses.

Is all AI writing slop?

No. Writing that someone reads, checks and makes their own isn’t slop, however it was drafted. Slop is the part nobody cared about.

Why do people dislike AI slop so much?

It wastes their time. A reply that could sit under any post tells the reader the sender didn’t read theirs.