I built an AI LinkedIn tool. Here's my honest take.
Yes, I'm biased. We built Tyashin — an AI tool that generates and publishes LinkedIn posts. So take everything I say with that context.
But I think my perspective is useful precisely because I've spent the last year thinking deeply about where AI content works, where it fails, and where the line between "helpful tool" and "inauthentic shortcut" actually sits.
The answer isn't a simple yes or no. It depends on how you use it.
The case against AI content (it's real)
Let me start with the criticism, because a lot of it is valid.
The flood of garbage. Since ChatGPT launched, LinkedIn has been flooded with AI-generated posts that all sound the same. You know the type: "5 lessons I learned about leadership" written in that painfully neutral, hedge-everything, use-every-buzzword tone. These posts are technically correct and completely soulless.
The trust problem. When readers suspect a post is AI-generated, they disengage. A study from the University of Zurich found that people rate AI-generated content as less trustworthy even when they can't reliably identify which content is AI vs. human. The perception alone is damaging.
The sameness. Most AI tools produce output that sounds like... every other AI tool. There's a "GPT voice" — polished, slightly formal, relentlessly upbeat, full of phrases like "it's worth noting" and "at the end of the day." It's recognizable. And recognition is the death of authentic content.
These are legitimate problems. And they're exactly why we built Tyashin the way we did. But I'll get to that.
The case for AI content (it's also real)
Here's the other side.
Most people don't post at all. The average professional has something valuable to say but never says it. They're too busy. They're intimidated by the blank page. They can't justify spending 45 minutes writing a post when they have a business to run. If AI helps them go from zero posts to three posts a week, that's a net positive for everyone — the creator, their audience, and the platform.
The alternative isn't "they'll write it themselves." That's the fallacy in most anti-AI arguments. The choice isn't between AI content and hand-crafted content. It's between AI-assisted content and no content at all. For most busy professionals, silence is the default.
AI as a draft partner, not a ghostwriter. There's a spectrum. On one end: copy-paste from ChatGPT without editing. On the other: use AI to generate a first draft, then rewrite, reshape, add your voice and examples. The second approach isn't inauthentic — it's just efficient. Professional authors use editors. Executives use speechwriters. This is the same principle.
Where the line actually is
After building in this space for a year, here's where I think the line sits:
AI is fine when:
- You provide the ideas, opinions, and experiences — AI handles the structure and polish
- You review and edit every post before it goes live (or at least have the option to)
- The AI is trained on your voice, not just generic "professional" tone
- The content is grounded in real knowledge — your expertise, real data, actual experiences
- You'd be comfortable saying "I used AI to help write this" if asked
AI crosses the line when:
- You're publishing content about experiences you never had
- The opinions in the post aren't yours — they're whatever the AI generated
- You couldn't explain or defend the post's argument in a conversation
- You're using it to fake expertise you don't have
- The content is indistinguishable from a thousand other AI posts
The key question isn't "was this written by AI?" It's "does this represent my actual thinking?" If the answer is yes, the tool you used to get it on the page is irrelevant.
How to use AI without sounding like AI
If you decide to use AI (and I think most professionals should), here's how to do it well:
1. Invest in the setup
Don't just type "write me a LinkedIn post about leadership." That's how you get garbage output. Instead, invest 20-30 minutes setting up your voice profile. Tell the AI how you speak, what words you never use, what opinions you hold strongly, what your audience cares about.
With Tyashin, this happens during onboarding — you define your voice, upload your knowledge bank, set your content preferences. The AI references all of this in every post. With ChatGPT, you'd need to include this context in every prompt (which is tedious but possible).
2. Always edit the output
Even the best AI tools produce drafts, not finished products. Read every post before it goes live. Ask yourself:
- Would I actually say this?
- Is there a specific example from my experience I could add?
- Does the hook feel like me or like a template?
- Are there any phrases that feel "AI-ish"?
Change anything that doesn't sound like you. Add your own examples. Remove any sentence you wouldn't say out loud. This 5-minute editing pass is the difference between AI-assisted content and AI-generated content.
3. Feed it your real ideas
The best AI-assisted posts start with a real idea from you. A conversation you had. A problem you solved. A opinion you've been chewing on. Give the AI the raw material and let it shape it into a post.
The worst AI-assisted posts start with "write me something about [broad topic]." There's no you in that prompt. The output will reflect that.
4. Use it for consistency, not creativity
AI's biggest value isn't writing brilliant posts. It's ensuring you post consistently. Three decent posts per week will outperform one brilliant post per month. Every time.
Use AI to handle the posts that would otherwise not exist. The Tuesday post you'd skip because you're in back-to-back meetings. The Thursday post you'd abandon because you can't find the right words. Those are the posts AI should handle.
Save the hand-crafted posts for when you have something truly personal or important to say. Maybe that's once a week. Maybe it's once a month. Both are fine.
What happens next
AI content tools are going to get better. Voice fidelity will improve. Content grounding will improve. The gap between AI-assisted and human-written will shrink.
At the same time, audiences will get better at detecting lazy AI content. The bar for what's "good enough" will rise. Tools that just slap a LinkedIn template on top of GPT output will stop working.
The winners will be the professionals who use AI as a force multiplier for their real expertise — not as a replacement for having something to say.
So should you use AI to write LinkedIn posts? Yes. But use it like a smart tool, not a crutch. Bring your ideas. Review the output. Keep it honest.
The world doesn't need more AI-generated content. It needs more of YOUR thinking — published consistently, structured well, and delivered to the people who need to hear it.
AI just makes that possible.