Using AI on LinkedIn without losing your voice

You can usually tell within two lines. The rhythm is too even, every sentence lands on the same note, and there is a triad where a normal person would have written one thing. The post is competent and completely anonymous.

That is the real risk of AI on LinkedIn. Not that it writes badly, but that it writes like the average of everyone who has ever posted. On a platform where your name is the asset, average is the one thing you cannot afford.

Feed it your words, not your topic

Most people give the model a subject and ask for a post. That is how you get the average. The alternative takes longer once and then keeps paying:

• Record yourself talking about the subject for ten minutes. Not writing, talking.
• Transcribe it, mistakes and all. The hesitations are where your voice actually lives.
• Ask the model to cut and organize that transcript, never to rewrite it.

One monthly conversation of this kind is enough to feed several weeks of content. The model becomes an editor rather than an author, which is the only role it is genuinely good at.

The forbidden list

A short list of banned habits does more for authenticity than any amount of prompt engineering. Ours:

• No rhetorical questions as an opening line.
• No sets of three when two would do. The model loves triads, people do not talk that way.
• No words you would not say out loud. If you have never said “leverage” to a colleague, it does not go in your post.
• No conclusion that summarizes what was just said. Stop when the point is made.

How to tell when it has drifted

Read the draft out loud. Not silently, out loud. Every sentence you stumble on is a sentence you would never have said. That test catches more drift than any tool, and it takes ninety seconds.

The second test is subtraction. Remove your name and photo from the post. If a colleague in your field could have published it word for word, it carries no voice, and it will build no reputation.

Where the line sits

Use AI for the work that is genuinely mechanical: sorting a list, reading a profile, cutting a transcript, catching a repetition, suggesting an angle you had not considered. Keep for yourself the things that make people trust you: the opinion, the example from last Tuesday, the thing you got wrong and what it taught you.

AI can make you faster. It cannot make you worth following. That part is still your job.