Is this LinkedIn post AI-generated? How to tell, and what to do about it
6 min read
You have read the post. Even pacing, a confident hook, three tidy lessons and a 'Thoughts?' at the end. Something feels off, and you want to know: did a person write this, or did a model?
The honest answer is that nobody can prove it. What you can do is look for the mechanical patterns machine writing tends to leave behind. This guide walks through them, and points you to a free checker that does it for you: Yolect's AI checker at yolect.com/tools/ai-detector.
Why no tool can prove a post was written by AI
Any detector that claims certainty is selling something. A careful editor can strip every machine pattern out of an AI draft, and a tired human writer can trip half of them on their own.
So treat any result — including ours — as a set of signals, not a verdict. The useful question is not 'was this AI?' but 'does this read like a person with something specific to say?'
The eight signals to look for
These are the patterns the free checker measures. Each one alone is weak evidence; several together are a fingerprint.
- Stock phrases: 'in today's fast-paced world', 'unlock', 'leverage', 'game-changer', 'delve', 'let that sink in'.
- Metronome rhythm: almost every sentence is the same length. People write in bursts.
- Suspiciously even paragraphs: every block is about the same size.
- The em-dash habit: several em-dashes in one short post. Most people rarely type the character.
- Nobody home: no 'I' or 'we' anywhere in a long post.
- Nothing concrete: no number, name, date or moment — nothing only the writer could know.
- Over-decoration: a stack of exclamation marks and emoji standing in for tone of voice.
- A formulaic ending: a prompted question like 'Thoughts?' or 'Agree?' bolted on for engagement.
How to check a post in ten seconds
Paste the post into the free AI checker at yolect.com/tools/ai-detector. It runs entirely in your browser — nothing is uploaded or stored — and gives a 0–100 score with a plain sentence for each signal it found.
The first three signals are open to everyone. A free account unlocks the full eight-signal report.
If your own post gets flagged
Every one of these signals is a habit you can fix. Break the even rhythm with one very short sentence. Cut the stock phrases. Add one detail from your actual week — a number, a customer's words, a decision you made.
The deeper fix is to stop starting from a blank prompt. A post built from your own voice note or rough thought already carries the specifics and the rhythm a model cannot invent. That is the workflow Yolect is built around.
Questions people ask
- Is the AI checker free?
- Yes. The score and first three signals are free with no account. A free account unlocks all eight signals; no card is needed.
- Does the checker store what I paste?
- No. The analysis runs in your browser, in code, with no upload and no AI model call.
- Does LinkedIn penalise AI-generated posts?
- LinkedIn has not published a rule that penalises AI-written posts as such. What readers do penalise is generic writing — they scroll past it. That is the real cost.