Yes, it's OK to use AI to write LinkedIn posts. AI is a professional writing tool, the same way code assistants are for developers and Figma is for designers. The line is authorship: the idea, the opinion and the story must be yours and true, and you must read and approve every post. Using AI to publish thoughts you don't hold, or anecdotes that never happened, is where it stops being OK.
Yes. It's OK to use AI to write your LinkedIn posts, and if you write professionally without using AI for anything, not even brainstorming or grammar, you're probably behind.
What matters is who owns the idea. If the opinion is yours, the story actually happened to you, and you read and approved the final post, then the post is yours, whichever tool helped you type it. If the AI came up with the point of view, or invented an anecdote about a coffee meeting you never had, that's where it stops being OK.
I run a two-person company where there is almost no task I do without AI. I also built a product that drafts posts for founders. So I have skin in this. I also have strong opinions about where AI writing goes wrong, because I've watched it go wrong up close.
Nobody shames developers for using AI
Think about how we treat other professions.
A developer who doesn't know how to use an AI coding assistant today is behind. Nobody calls the code inauthentic because Claude wrote half of it. A designer who refuses to use Figma and sketches every screen by hand is slow, and nobody praises them for it. Nobody wants a designer to hand-draw a SaaS product.
Writing is the odd one out. A founder uses AI to help draft a post and suddenly someone in the comments calls it fake. People screenshot "ChatGPT-sounding" posts to mock them. Some founders I talk to avoid AI entirely because they're afraid of being caught.
I understand where it comes from. Many people who can't write use AI to generate everything, publish the first output, and flood the feed with low-quality posts. That flood has made writing with AI, in general, feel like something to be ashamed of.
In a few years, I think this will look as strange as shaming someone for using spell check. Nobody will expect a professional writer to work without AI amplifying what they do. So I'd stop asking whether someone used AI and start asking a better question: is the content good, or is it lazy?
Where the real line is: authorship of the idea
Here is how I'd separate acceptable from not acceptable. What decides it is what the tool contributed.
| AI's role | OK? | Why |
|---|---|---|
| Fixing grammar, tightening sentences | Yes | Editing |
| Turning your voice memo or notes into a draft | Yes | Your idea, faster |
| Suggesting hooks, structure, a stronger ending | Yes | Craft help, you choose |
| Brainstorming angles on a topic you know well | Yes | You judge which angles are true to you |
| Drafting a post from a one-line prompt, then posting unread | Risky | Nobody owns the idea |
| Stating opinions you don't actually hold | No | That's someone else's post under your name |
| Inventing a story, client, or conversation | No | That's fabrication |
The top of the table is where most founders should live. You bring the raw material: what happened on a customer call, a strong reaction to a piece of news, a lesson from a mistake. The AI helps you get it into shape. That's the workflow I use myself, usually starting from voice memos I record between meetings.
The bottom of the table is the problem. And the worst item on it deserves its own section.
The one AI mistake that burns trust: made-up stories
When we started drafting posts for founders, one failure mode stood out. I now consider it the most dangerous one in AI content.
AI is very good at writing plausible anecdotes. "A founder told me over coffee last week that..." It reads perfectly. It has a setting, a turn, a lesson. And the conversation never happened.
When the AI proposes a specific story like that, the client's reaction is predictable: "No, that's not my story." They feel disappointed, and they're right. A CEO can't post about a meeting that never happened. Their own team would read it and know.
So we set a hard rule. If there's no factual basis that a story happened, the AI doesn't write the story. It works from principles and points of view the person has already expressed, in past posts or in their own notes, and leaves the stories to them.
If you use a general-purpose chatbot to write your posts, watch for this closely. Every time a draft contains a specific scene, a named-sounding person, or a number you didn't provide, ask yourself where it came from. If the answer is "the AI," delete it or replace it with something that actually happened.
Using AI for LinkedIn posts well: a simple workflow
This is the workflow I'd recommend to any founder who wants to use AI without losing their voice:
- Start from something real. A voice memo, meeting notes, a Slack message you wrote to your team, a comment you left on someone else's post. Never a blank prompt like "write a post about leadership."
- Ask for a draft, then argue with it. Tell it what's wrong. Push back on the weak points. In my experience, AI advice and drafts get useful after a few rounds of challenging them. The first draft blends good ideas with bad ones, and you need your own judgment to tell them apart.
- Cut anything you didn't say or live. Invented anecdotes, borrowed opinions, statistics you can't source.
- Read it out loud. If you wouldn't say it that way to a colleague, rewrite it.
- Check it before you post. Paste it into an AI-sounding post checker to catch the patterns readers now spot instantly.
- Approve it yourself. Your name is on it. Nothing goes out that you haven't read.
Step 2 matters more than people think. I've seen experienced businesspeople stay very patient with AI, going round after round, because some of what it gives them is really good and advances their thinking. Someone without that experience takes the first answer and ends up publishing something generic or wrong. AI makes experienced people faster. It doesn't give inexperienced people judgment.
What I learned watching 400 people try ChatGPT
Back in January 2023, when ChatGPT was the only tool anyone was talking about, I opened a Slack channel at the company I was working at and called it ChatGPT. The opening message said, roughly: here is an amazing new technology, it can change how we work, let's use this channel to share use cases where it helps.
About 100 people out of 400 joined. That was a good start. Within a short time the channel became a place to post jokes about ChatGPT hallucinating.
People were too busy to learn a new tool, and "invest time now to save time later" didn't land. In my experience it was only around mid-2024 that the team really started to shift.
Now compare that to my current company, which is two people. There's almost nothing I do without AI. All our code is written with Claude. Every meeting is recorded and summarized. After sales calls I get coaching from AI. Tasks land in our project tool automatically. Our website and even our logo were built with AI. When you start clean, it's easy to use AI for everything, because you understand there's no other choice.
The founders I see hesitating to use AI for LinkedIn remind me of that Slack channel. They're focused on the jokes and the failures, while the people who learned to use it well are already posting consistently in their own voice. The narrative I keep coming back to: the best leaders in the AI age are hands-on. You can't lead what you've never touched, and that includes the tools your team is already using.
Does using AI hurt you on LinkedIn?
Two separate worries hide in this question.
The first is the algorithm. I wrote a longer piece on whether AI content hurts your LinkedIn reach. The short version from what I've seen: generic content performs badly, whoever wrote it. A post with a real point of view and a real story does fine, whoever typed it.
The second is people. Readers are getting good at spotting default AI style: the same rhythm, the same phrases, the same tidy three-point structure. When they spot it, they scroll, and some of them trust you a little less. That's a quality problem, and it's the one worth your attention. Most AI posts sound like AI because something is missing from the input: your specifics, your opinions, your stories.
So using AI on its own doesn't hurt you. The real risk is posting what the AI thinks under your name.