LinkedIn 'Seems like AI slop' button: what AI video posters change

LinkedIn added a way for members to mark a post as "seems like AI slop", and the signal goes to the feed team to tune its models. An AI video is not banned on LinkedIn by that button. A video that looks generic, recycled or made in bulk is what gets flagged, so the fix is in the brief and the cut, not in hiding that you used AI.

The button was announced on 2026-07-30 by LinkedIn's chief product officer, Hari Srinivasan, as reported by TechCrunch. That is a press report, not a LinkedIn page, so treat the details as reported. The page we could read from LinkedIn itself is its engineering write-up on feed quality.

What was reported, and what LinkedIn wrote

Each row says where it comes from, so you can check it before you build a plan on it.

What an AI video post can change

LinkedIn's own line, as quoted by TechCrunch, is that people come to connect with real people and their real perspectives. For a brand or a creator that means the clip needs something only you have:

• A point of view in the first two seconds, written by a person, not a default prompt.
• A real fact on screen: a product detail, a process step, a customer outcome you can stand behind.
• A different brief per video. Ten clips from one prompt with one word swapped read as bulk.
• A named presenter or company behind the post, so a reader knows who is speaking.

How to do each step with Sume

Sume does not rank or score slop, and it cannot tell you what LinkedIn's classifiers think. It can make the steps above repeatable.

• Generate with a specific brief per clip through POST /v1/videos. Pin the first frame with frame_images when the opening has to be your real product or place.
• Burn your own facts as text with video captions: send cues with text, start and end so the on-screen words are yours, not a speech-to-text guess. A standalone caption job is $0.20 for a video of up to 60 seconds.
• Read the finished file with video inspect and transcribe: true. The speech-to-text rate is $0.01 per audio minute. Read the transcript aloud: if no sentence in it carries a fact or an opinion, cut the video.

Checklist before you post

Run this once per video. It takes a minute, and it is the part LinkedIn's classifiers and its readers both react to.

• Does the first line say something a competitor's video would not?
• Is every number, name and claim on screen true and yours?
• Is the post under a named person or company page?
• Would you post it if it had been shot on a phone? If not, rework the brief.
• Did you check the platform's rules on disclosure? See the disclosure checklist by platform.

What we could not confirm

We did not find a LinkedIn page that lists how the button changes reach, so this post does not state a reach effect. We also cannot say whether any one AI video is classified as slop. The same question for YouTube is in the Shorts originality checklist, and the file limits for a LinkedIn video ad are in the 4:5 spec check.