What did LinkedIn announce on 20 May 2026?
LinkedIn said it would show generic AI content to fewer people. On 20 May 2026, Laura Lorenzetti, executive editor of LinkedIn's global editorial team (per their LinkedIn profile), published Keeping conversations real on LinkedIn. LinkedIn's Pressroom republished it on 4 June 2026. It says LinkedIn is cracking down on automation tools and dialling back generic content, using systems built with its editorial team. Those systems are trained to tell content that adds perspective, context or expertise apart from content that feels generic or repetitive, even when it reads as polished.
The post is clear that AI help is allowed: "It's ok to use AI to help you write, but your posts and comments need to represent your voice and your perspectives."
Two numbers anchor the announcement. LinkedIn says that in initial testing it identified generic content correctly 94% of the time. It also says members can now filter for its more than 100 million verified members in comments and feed conversations, as well as in profile views and job applications.
What does LinkedIn now hold back?
Four kinds of content, plus fake accounts. The table pairs each with what the post says happens to it.
| What LinkedIn targets | How the post describes it | What happens |
|---|---|---|
| Generic AI posts | Content that appears to be AI-generated and lacks a clear perspective | Less likely to be widely distributed beyond the poster's immediate network |
| Repetitive posts | Content that feels generic or repetitive, even if it looks polished | Less likely to spread; you'll see less of it from outside your network |
| Automated comments | Comments created and posted at scale with automation tools, with little or no human involvement | LinkedIn's systems will recognise them and take action; the post doesn't say what action |
| Restating replies | Responses that repeat the original post without adding anything new | Same as automated comments |
| Bots and fake AI profiles | Bots and fake AI profiles that ruin genuine engagement | LinkedIn's teams work to remove them; readers can filter for verified members |
The line that matters for anyone posting with AI is that such content "is less likely to be widely distributed beyond a person's immediate network". The post says nothing about reach inside your immediate network. What it limits is reach into new audiences.
What hasn't LinkedIn said?
Quite a lot. The announcement is short, and these questions are left open in its text, as published on 20 May 2026:
- How the 94% was measured. The post gives no method and no figure for how often posts with a real point of view are flagged by mistake.
- Whether you are told. It mentions no notice to members whose posts are held back, and no way to appeal.
- What "immediate network" means. It doesn't say whether that is your connections, your followers or both.
- Whether good posts gain. It says generic content will travel less. It does not promise more reach for thoughtful posts.
- Whether messages are covered. It covers posts, comments and replies. In the comments under the post, one reader asked when the same approach would reach InMail, and the post itself says nothing about messages.
Plan around what is stated and treat the rest as unknown until LinkedIn says more.
What still gets reach on LinkedIn?
LinkedIn's post points to content that adds perspective, context or expertise. Those are the three qualities it says LinkedIn's systems learn to identify. In practice, a post needs something the reader could not get from a chatbot prompt:
- A position. You disagree with common advice and say why.
- Context only you have. A number from your own business, a decision you made, a mistake and what it cost.
- Expertise with the working shown. A method, a checklist or a before-and-after from your own work.
- Comments that add. A fact, an example or a question the original post didn't cover.
People are part of the loop too. In a 5 March 2026 post, Lorenzetti described LinkedIn's editorial team as over 150 editors and creator managers in 12 countries and seven languages, who work with engineering on what quality content looks like.
For post structures that start from your own material, see our 12 LinkedIn post templates.
Before and after: a generic post and one with a point of view
The difference is specific detail that only the author could know. Both examples below are written by us for a fictional founder of a 12-person accounting firm. The first is the kind of post LinkedIn describes as lacking a clear perspective; the second covers the same topic with a point of view.
What changed between the two:
- A specific event and month instead of a trend.
- A number the reader can weigh: 40 clients, 6 hours a week.
- A decision with a trade-off: calls over growth.
- A mistake, stated plainly, with what to do instead.
None of it needs to be long. Every detail in the second post is something an AI tool could not have known without the founder telling it.
What about comments and auto-engagement?
Automated comments are now a named target. The post says LinkedIn's systems will act on comments posted at scale through automation tools with little or no human involvement, and on replies that only restate the post. LinkedIn's help page on automated activity, checked 20 October 2026, says it doesn't allow third-party software or browser extensions that automate activity on its website.
Both are our examples for the same fictional founder. The second adds a method and asks a real question: something the original post didn't already say. For more structures like it, see our 30 LinkedIn comment templates.
If a tool comments for you, read each comment before it posts. If you use Siela's LinkedIn Presence agent, set publishing to "show me every post first" and engagement to "show me first". Each draft is written in your voice and checked by a second AI reviewer, and your read is where your point of view goes in. Reviewing drafts does not change LinkedIn's help-page rule on third-party automation tools; our post on whether LinkedIn automation is safe covers that risk.
How to use AI for research and drafts and keep your perspective
Let AI do the research and the first draft, and keep the point, the proof and the opening line for yourself.
Write your point first
Before you open any AI tool, write one or two sentences: what you believe, and the proof from your own work. If you can't write it, you don't have a post yet.
Use AI for research
Ask for data, sources and the strongest argument against your point. Open every link and keep only what you can confirm.
Draft from your notes
Give the AI your point, your numbers and your example, and ask for a draft in plain words. A one-line prompt such as "write a post about AI in accounting" gives the AI nothing of yours to work with.
Write the opening yourself
Put the specific detail first: the number, the decision or the mistake. Rewrite the first two lines by hand.
Cut the stock lines
Delete anything any company could have posted: "the future belongs to", "here's the thing", "let that sink in", a closing "Agree?". If a sentence fits every post in your feed, it goes.
Add one thing only you know
A number, a customer detail you have permission to share, or the decision you would make differently. Then post.
What to do this week
Audit what you already post, switch off unattended commenting, and write from your own material. A checklist:
- Read your last five posts. Mark any that someone else in your industry could have written word for word. That is the content LinkedIn now says it holds back.
- Turn off unattended commenting. Any tool that posts comments at scale without your review matches what the May post names as a target.
- Write five one-line points from this month's work. A number, a decision, a mistake, a disagreement and a customer question.
- Turn two of them into posts this week with the steps above.
- Get verified if LinkedIn offers it to you. Readers can now filter comments and feed conversations to verified members.
- Track saves and followers gained per post. Lorenzetti's March 2026 post says LinkedIn's editorial team shaped content analytics that include both. They show whether readers found a post worth keeping or the author worth following.
Frequently asked questions
Does LinkedIn reduce the reach of AI-generated content?
LinkedIn says content that appears AI-generated and lacks a clear perspective is less likely to be widely distributed beyond the poster's immediate network. It announced this on 20 May 2026. Using AI to help you write is allowed; LinkedIn says your posts and comments need to represent your own voice and perspective.
How does LinkedIn detect AI slop?
LinkedIn says it uses systems built with its editorial team, trained to recognise content that adds perspective, context or expertise and content that feels generic or repetitive. In initial testing it identified generic content correctly 94% of the time. The announcement does not say how that figure was measured or how often good posts are wrongly flagged.
Are automated LinkedIn comments allowed?
LinkedIn's help centre says it does not allow third-party software or browser extensions that automate activity on its website. Its May 2026 post adds that its systems now act on comments posted at scale through automation tools with little or no human involvement, and on replies that only restate the post.
Will my connections still see posts written with AI?
LinkedIn's post says generic AI content is less likely to be widely distributed beyond a person's immediate network. It does not say the content is hidden from that network, so the main loss is reach into new audiences. Content that adds perspective, context or expertise is what LinkedIn says its systems look for.
Can I still use AI to write LinkedIn posts?
Yes. LinkedIn's May 2026 post says it is fine to use AI to help you write, as long as your posts represent your own voice and perspective. Use AI for research and first drafts, then add your own position, numbers and examples before you post.
What is LinkedIn's verified member filter?
LinkedIn says members can filter for its more than 100 million verified members across profile views, job applications, and comments and conversations in the feed. The feed and comment filters were announced in its May 2026 post on cutting bots, fake AI profiles and generic content.
Sources
- Laura Lorenzetti, LinkedIn: Keeping conversations real on LinkedIn (published 20 May 2026)
- Laura Lorenzetti, LinkedIn profile (checked 20 Oct 2026)
- Laura Lorenzetti, LinkedIn: Meet the humans that help shape LinkedIn's feed everyday (published 5 Mar 2026)
- LinkedIn Help: Automated activity on LinkedIn (checked 20 Oct 2026)
- Siela: 12 LinkedIn post templates (checked 20 Oct 2026)
- Siela: 30 LinkedIn comment templates (checked 20 Oct 2026)