How to spot a ChatGPT-written LinkedIn post before you publish

Almost nobody scrolling LinkedIn thinks "this was written by AI." What they think is closer to "I've read this before," and then they keep scrolling. You never find out, because nothing happened. No comment, no reply, no note in your inbox two weeks later saying the post was the reason someone got in touch.

That reaction is worth taking seriously, because it is not really about AI. It is about familiarity. Language models are trained to produce the most probable next sentence, which means they produce the sentence you have already read a hundred times. The writing is competent and completely forgettable, and forgettable is the one thing a post cannot afford when it is competing with everything else in a feed.

The tells, and why each one costs you readers

Six patterns show up again and again. Each has a reason behind it, which matters, because once you know the reason you can fix the cause instead of playing whack-a-mole with phrasing.

• Openers that announce a lesson: "Here's what I learned about leadership." It tells the reader the conclusion before they have any reason to care about it. A model writes this because summaries are safe. A person writes the moment instead, and the moment is what earns the second line.
• Perfectly even rhythm, where every sentence lands at roughly the same length. Human writing is lumpy. A long, slightly overpacked sentence followed by three words. That unevenness is what makes text feel spoken rather than assembled.
• Abstract nouns doing all the work: alignment, journey, mindset, unlock, leverage. They are frictionless to generate because they fit anywhere, and that is exactly the problem. A word that fits any post carries no information about yours.
• Endings that restate the post. Models close loops because closing loops looks tidy. Readers have already read the post, so the restatement is dead space where the strongest line should be.
• No proper nouns and no numbers. No client type, no city, no week, no price, no failure. Nothing in the post could only have come from your life, which is the definition of generic.
• Even-handedness about everything. Real expertise means you have opinions that cost you something. A draft that agrees with all sides has no author.

A pass that takes about three minutes

Read the post out loud. This sounds like advice from a writing class, but it works for a mechanical reason: your mouth refuses sentence shapes your eye will happily skim past. Anywhere you stumble, or anywhere you would not say the sentence to a colleague across a table, rewrite it in the words you would actually use.

Then delete the first line and read it again. In most AI drafts the post is better without it, because the first line is setup rather than content. If deleting it loses nothing, it was costing you the only two lines the feed shows before "see more."

Last, add one concrete detail to each paragraph. A number, a name, a date, a client type, a mistake with a price attached. Specificity is the thing AI text is structurally missing, and it is also the fastest thing to add back, because you already know the details. You just left them out while you were trying to sound professional.

Why the fix works

Specific writing is harder to fake and easier to trust. When a reader sees "a founder emailed me a deck at 1am on a Sunday," they can picture it, and being able to picture something is most of what people mean when they say writing feels real. Generic writing asks the reader to take your word for it. Specific writing hands them the evidence and lets them draw the conclusion themselves, which is also why it gets argued with in the comments, and comments are what put the post in front of people who do not follow you.

Where Post Check fits

Post Check runs this pass on the draft before you publish. It marks the lines that read as generic, says why each one was flagged, and shows the phrasing that matches how you have written in the past. You keep the final call on every line, which matters, because half of good writing is knowing when to break your own rules.