Tech and AI
More than a million people have used LinkedIn's button for reporting posts that read as machine written
By Staff Writer | 22 August 2026
The reporting option was added on 30 July and sits in the menu on every post. Hari Srinivasan, the platform's chief product officer for its ecosystem, said on 20 August that more than a million people have used it, and that material the platform classifies as low effort machine writing is now getting about 40 per cent fewer views than it did a few weeks ago.
The option itself is plainly worded. It is reached from the three dot menu on a post and it says the post seems like machine generated filler. There is no appeal, no penalty stated and no removal; what a report feeds is a classifier, and what the classifier changes is distribution.
The measurement behind it
The button arrived weeks after an outside detection service reported that 41 per cent of longform posts on the platform were flagged as fully machine generated. The platform did not accept or contest that figure. It shipped the button, new classifiers, and the removal of a feature of its own that had offered to rewrite a user's post with artificial intelligence.
AI slop is a top priority for all of us
Hari Srinivasan, chief product officer for the LinkedIn ecosystem, at the launch on 30 July
The 40 per cent reduction is the platform's own account of its own classification, which is worth stating plainly. It measures views of material the platform has decided reads as machine written. It does not measure how much machine written material is being posted, and it does not establish that any individual post was in fact machine written.
The message to the author
The change now being added is the one professionals will notice. An author whose post attracts reports will be told that some members thought it looked machine written. There is no accusation and no sanction attached, but the platform is putting a judgement about the manner of a person's writing in front of them.
We approached this assuming good intent; I know I'm increasingly conscious on how to not sound like AI & the goal is to provide helpful feedback
Hari Srinivasan, chief product officer for the LinkedIn ecosystem, 20 August
What it means for people who post for work
Consultants, surveyors, engineers and lawyers use this platform as a shop window, and a good deal of what they post is drafted with help. The line being drawn is not between assisted and unassisted writing. It is between writing that carries a point and writing that reads as filler, judged by other members and then by a classifier trained on what they flagged.
The practical effect is that a certain register now costs reach: the tidy three part structure, the rhetorical question opening, the closing invitation to share your thoughts below. None of that is banned. It simply travels less far than it did in July, and the platform has published a number for how much less.
Earlier this year the same platform said it would act against comments produced at scale with little or no human involvement. The reporting button extends that from comments to posts, and hands the first judgement to the readers.