LinkedIn has a serious problem. And if you're using LinkedIn, you know what it is. The worst A.I.-generated posts in the land are infiltrating news feeds and ruining the vibe. LinkedIn doesn't know if it loves or hates AI. And LinkedIn users are losing trust .

A survey found that 52 % of consumers are concerned about brands sharing undisclosed AI content. Nearly half of U.S. consumers say generative AI has made media content worse . Audiences have figured out what AI-written copy sounds like, and they are punishing brands that publish it.

Personal LinkedIn posts written by a human, on a personal account, with a real opinion attached to a real face, is the format the algorithm is rewarding. If you have been outsourcing your LinkedIn to a tool that writes posts from a prompt, the data is now against you.

What the LinkedIn algorithm started doing in May

The May 2026 LinkedIn algorithm update changed two things. One, reward comments more than likes. Two, it penalises posts where average dwell time falls below ten seconds. AI-written posts lose on both counts. They generate superficial likes from connections who scroll past. They lose comments because the content has no opinion to disagree with. They lose dwell time because the post sounds like every other AI post on the feed.

Posts with a direct question get 77% more comments than the average post. Posts with a personal anecdote get four times the dwell time. The platform is paying you to be a person.

The update also changed how the algorithm scores repeat posters. Accounts that post daily with the same structure get throttled within two weeks. Accounts that post three times a week with varied formats get a steady reach lift. Consistency in 2026 looks like varied formats posted three times a week.

What audiences are pattern-matching

Three signals trigger the AI tag in a reader's brain. The first is generic openers. The sweeping setup about modern work, the rhetorical question, the cliched stat about Gen Z. “Most Gen Z workers aren’t looking for long term.” “What if you could build the business of your dreams.” Yawn. If you start a post like this, the reader knows the next sentence before reading it.

The second is smooth transitions. The connector words that string two paragraphs together without any thinking. AI writing connects every sentence with a smooth transition word. “Here’s what this means.” “But you didn’t know this…” “Meanwhile…” That smoothness is the giveaway.

The third is symmetrical structure. A neat opener, three balanced points, a tidy close. If a post reads like it was assembled from a template, it was. The reader does not need three paragraphs to spot a fake. The pattern is recognised in two seconds and they scroll on.

The advantage for personal brand builders

You have a stronger hand here than you realise. You have built a business. You have created success for clients. You have watched a launch go wrong and figured out why. You have opinions that came from doing the work fifty times. That is the source material AI cannot fake. Not yet. Probably not ever in a way the algorithm will not catch.

Publish the things you would never let an AI write. The story about the client you fired. The number on the deal that fell through. The specific reason you changed your pricing. These posts feel slightly risky to publish. And posts that feel uncomfortable to share are the posts the algorithm rewards.

The fear of publishing the real thing is the only barrier between you and the audience that already wants to hear it. Founders who push past that fear once can push past it weekly. The compounding effect on visibility is big.

A test you can run this week on LinkedIn

Write your next five posts to a different brief. Pick one moment from the last seven days. A meeting that went sideways. A piece of feedback that landed. A number you saw on a dashboard. Write three sentences about what happened, three sentences about what it changed in your thinking, and one line on what you are doing differently next week. Frame it as unsolicited advice to your dream client.

No headers. No bullet points. No hooks. No emojis. Sentence case throughout. Get a bee in your bonnet and go on a rant. Post once a day for five days. Then check the comments-per-post ratio. The data will tell you whether the new approach works for your audience.

Founders see the comment number rise within the first week. Reach lags by a fortnight. By month two the inbound starts to change. Less spam, more invitations to speak, podcast, or partner. That’s the good stuff.

The bigger play of human LinkedIn posts

The AI authenticity backlash is real. There’s a ceiling on what generic content can earn. The amount of AI-written copy is increasing. The signal-to-noise ratio favours anyone willing to publish the real thing.

Do this for the lurkers, not the likes. Don’t forget that the buyers who hire you in 2026 are reading every post you publish, even if they never engage. They are looking for proof that you think for yourself. A real opinion, well argued, on your own profile, is a valuable asset.

Write fewer posts. Make each one great. Say the thing you have been editing out. Do this for the next twelve months, own categories you were nowhere near in 2023.

Get my free workshop on how to write with AI .