LinkedIn shipped a button today that says "Seems like AI slop." That is the actual label. Not "low quality," not "inauthentic content," not some committee phrasing. Slop. The most AI-flooded network on the internet decided the honest word was the marketing word, and I have to give them that one.
Then they did the part that actually matters, which nobody is leading with: they pulled their own AI writing button.
What changed on Thursday
TechCrunch's Sarah Perez reported the announcement on July 30, 2026, sourced to a LinkedIn post from Chief Product Officer Hari Srinivasan. Engadget's Karissa Bell covered the same update the same afternoon. Between the two reports, here is what LinkedIn says it is doing.
You can now flag a post as "seems like AI slop." That flag does two jobs. It feeds LinkedIn's classifiers so they get better at spotting the stuff, and it privately tells the person who posted, through their own analytics dashboard, that readers felt the post "may have come off as inauthentic or heavy use of AI." A LinkedIn spokesperson told Engadget flagged posts also get reduced reach, working roughly like a "not interested" signal.
New classifiers are supposed to cut how much AI slop shows up in recommendations from outside your network. LinkedIn is expanding profile and page verification. You can block comments from company pages you are done hearing from.
And the "enhance your post" feature, the one LinkedIn put in front of every user, is gone. In its place is a proofreader that, in Srinivasan's words, "proofreads your words, but does not change your voice."
The number that explains the panic
Perez and Bell both point at the same context, and it is ugly. Engadget cites a study from AI detection firm Pangram finding LinkedIn was the most AI-saturated social platform, with more than 40 percent of longform posts flagged as fully AI-generated.
Two out of five long posts. Written by a machine. On the network people use to decide whether to hire you.
I want to flag that number correctly, because I care about this: that is a detection company's estimate, and detection companies sell detection. AI-detection tools have a documented false-positive problem. Take 40 percent as directionally right and precisely unproven.
The bot side is LinkedIn's own admission and I weigh it more heavily. Srinivasan said the company is blocking hundreds of thousands of automated comment attempts, and millions to billions of other automation attempts over recent months, depending on which report you read — TechCrunch says "hundreds of thousands" of comment attempts daily and millions of automation attempts over a couple of months; Engadget frames it as hundreds of thousands of automated comments caught and billions of posting attempts blocked in the last few months. Either way, the scale is industrial and the platform is saying so out loud.
Why pulling their own button is the real news
Every platform right now wants credit for fighting slop while shipping the tools that generate it. LinkedIn put "rewrite with AI" under every post box, then acted surprised when the feed filled with the same five-paragraph humblebrag in the same voice.
Killing that button is LinkedIn admitting it was part of the supply. That is rarer than a moderation feature and it costs them something real, because engagement teams love a button that makes posting easier.
The verdict: the reporting button is a signal-collection play, the proofreader swap is the actual policy change.
What this means if you post for work
I write in my own voice on purpose, which is why I have opinions here.
Do not stop using AI. Stop using it as the author. The proofreader-not-rewriter line LinkedIn drew is the correct line for anybody publishing under their own name. Let a model catch your typos and your run-on sentences. Do not let it decide what you think, because that is the thing readers can smell and now the thing they can report with one click.
Expect false flags. A reduced-reach penalty triggered by crowd reports will land on people who simply write formally, or who write in a second language, or who use an em dash. If your reach drops for no reason you can trace, check your dashboard for that private nudge before you rewrite your whole strategy.
If you run automated posting, read the room. LinkedIn is blocking automation at scale and has said so publicly. Anything that posts or comments on your behalf without a human in the loop is now aiming at a moving target. This is the same lesson I keep hitting with autonomous tooling in general: an agent acting under your identity is your liability, and you should know exactly what it is permitted to do before it runs unattended. I wrote the checklist version of that in how to audit your AI agent permissions, and the difference between a tool that answers and a tool that acts on your behalf is the whole subject of AI agents versus chatbots.
The part I do not think will work
A crowd-sourced report button on a network of 1 billion accounts is a rounding error against the volume LinkedIn just described. Engadget says it plainly: given the sheer amount of AI material, it is unclear whether any of this changes the trend.
I would go further. Reports are a slow, noisy, human-rate-limited signal aimed at a problem that generates content at machine rate. The classifiers might scale. The button will not. The button is mostly a way to get labeled training data from users for free, which is a smart move and a small one.
What would actually work is the incentive side: kill reach for accounts with a sustained slop pattern, publish the rule, and eat the complaints. LinkedIn has not gone there.
Where this fits
This is not an isolated move. TechCrunch notes Substack shipped an AI-identification tool last week through a partnership with Pangram, that Pangram just raised $9 million, that Digg shut down its Reddit competitor in March over bots, and that Cloudflare now measures more bot traffic on the web than human requests.
That last one is the story under the story. The open web is being outnumbered, and the platforms are shipping buttons.
My verdict: LinkedIn did one genuinely good thing today and one thing that will look good in a press cycle. The good thing is removing their own slop generator. Use the button when you see the stuff, keep writing in your own voice, and do not expect your feed to feel human again by Monday. Ya know?
Sources: TechCrunch on the "seems like AI slop" button, Engadget on the reporting tool and the removed AI writing features, CNET's coverage.