These four pointers can help you use AI to produce thought leadership that sounds human—and even like you.

---

I recently saw an article “written” by an attorney in a legal trade publication that made me want to yell. Here’s how it started:

There was a time when these phone calls were easy.

Someone would tell me they had settled an employment dispute. The documents had been signed. Payment was supposed to arrive weeks ago. It never did. Could I help enforce the agreement?

The next step was almost automatic.

“Send me the settlement agreement.”

I still ask for it.

I just don’t assume it tells the whole story anymore.

The article continued for another 550ish words. When I ran it through two separate AI detectors, they reported it was 100% AI-written. If you read the rest of the article, you’d agree.

These fake first-person articles drive me crazy, and they do more to hurt an “author’s” reputation than to help it. 

That’s because an article like that doesn’t express the author’s thoughts. The words attributed to the author aren’t their words; they’re an AI platform’s words that the author is claiming as their own.

Any attorney practicing in the same area could have written that article-or prompted an AI platform similarly to how the author did and gotten the same output. 

That’s the opposite of how your thought leadership should sound.

Your thought leadership is a marketing and branding tool that carries your name. By doing so, you claim ownership of the ideas and their expression, implying both are yours. The content should sound like you. Handing the virtual pen to AI when you’re writing thought leadership, to the point that the material doesn’t reflect your thoughts, experience, or voice, isn’t how you make friends or influence people.

---

---

Deferring to AI when you write is not the same as working with a ghostwriter

Let me get something out of the way. Having AI write your thought leadership for you is NOT ghostwriting.

When ghostwriters or speechwriters work for a client, the thinking starts with the client, and the writer’s words and finished product are based on the client’s words.

Ghostwriters mold the raw clay of their clients’ expressions of their experience, insights, guidance, and opinions into coherent, compelling, and distinctly theirs written or spoken words. Those words give form to their clients’ thinking and bring their voices to life.

In contrast, when someone types “Write a first-person article about lessons learned from a litigator’s first 20 years of practice” into ChatGPT or Claude, there’s nothing of theirs to shape. The AI platform will fill the page with whatever content it can find based on its training data.

The reputational issue caused by relying on AI to “write” your thought leadership

Based on my (very) informal research, few people who read legal thought leadership run it through AI detectors. I’m a crazy person who ghostwrites thought leadership for a living, so I run one or two on every thought leadership article I read. However, even people who don’t use AI detectors are getting better at noticing when a piece of writing carries someone’s name but was written by AI. 

They can tell the hallmarks of AI-written thought leadership. The content lacks personality; it feels and sounds hollow, and the lessons “learned” are ones anyone could have learned and communicated. There’s no connection between the person supposedly sharing those thoughts and opinions and the thoughts and opinions themselves. 

That’s a problem because the content we produce under our names shapes our target audience’s perceptions of us. When you put your name on a hollow, personality-less, intellectually empty piece of content like the one I quoted above, you’re going to make someone who reads it conclude your content isn’t worth their time.

And if your content isn’t worth their time in that moment, your content and your thinking might never be worth their time. That’s not good, because the chances of them hiring you after coming to that conclusion are slim to none.

How to use AI to help produce content that doesn’t sound like AI

That said, AI can be a helpful partner in your thought leadership efforts. It can help you brainstorm and refine topics, outline how you’ll present your thoughts, and stress-test your ideas by pointing out weaknesses in your arguments and counterarguments you didn’t address or consider.

And, if you know what you’re doing, you CAN use it to draft thought leadership that sounds like you and avoids harming your personal brand in the ways I mentioned above.

Here are four pointers to help you do so.

First, start with your own thoughts

It sounds obvious, but based on the content I’ve read from attorneys, it apparently isn’t: For thought leadership to sound like you, you have to provide most of the core thesis, supporting arguments, and context, so the content reflects your thoughts. 

When working with clients, before we write a word, we talk through a topic with them and draw out from them their thesis, arguments, context, view of the world, and anything else that allows us to craft an article that is true to them-and relevant and compelling to their target audience.

When you’re writing for yourself, I highly recommend recording a voice memo where you talk through your topic for five or 10 minutes before you touch a keyboard. Use the time to explore different angles and perspectives, discuss your thesis, arguments, context, and worldview, and decide how you’ll express them in a way that best serves your target audience.

That session will produce more interesting, authentic content than anything AI can generate on its own. By giving yourself some time to talk to yourself, your creative juices will start flowing.

When you’re done with your spiel, you could feed the audio or transcript into AI to see if it tightens your language or rearranges your thoughts in a way that still sounds like you, carries the message you’re trying to convey, and is more logically organized. But obviously, it’s relying on your words and thoughts-not on the words and thoughts it ingested during its training and is now regurgitating.

Second, sprinkle in specificity

The second way to make AI-assisted thought leadership sound like you is to include specific examples from your real-world experience. This is showing your work like you did in grade school with long division and other math. Sprinkle in examples of situations you’ve helped clients navigate, while being mindful of privilege, confidentiality, and other ethical issues.

Those examples show readers the content comes from your lived experience and proactively rebuts the concern that what they’re reading is just bland, undifferentiated AI slop. You might not be able to do this every time, depending on the topic you’re covering, but you’ll have plenty of opportunities.

I do this frequently with clients, particularly when we’re helping them write a “best practices” or “misconceptions” article that’s based on their experience handling certain kinds of matters. I ask them to recount real-life examples of when they employed a “best practice” or encountered a misconception. Even if you’re writing about the implications of a court decision or a regulatory agency’s action, you could share real client examples that would have been different if they’d happened after the court decision came down or the regulatory action was in effect.

In the article I mentioned at the beginning, the “author” inserted supposedly firsthand material, but because the entire article appears to be AI-written, that firsthand material seems fake. That shows that you can’t just add a single real-life example or rely on AI to produce a fake one. Instead, add enough specificity to push a reader over the hump from “Huh, this could be AI” to “This was clearly written by a human because they’re talking about several real-life examples.”

Third, rub people the wrong way

Want your AI-assisted thought leadership to not sound like generic AI output? Say something that people could disagree with.

As you’ve probably experienced, AI platforms are often agreeable by default. When you ask them questions, you’ll often get a balanced overview. That’s helpful for working through a problem and brainstorming, but it hurts your chances of making your content memorable and shareable.

When you use AI to craft your thought leadership, you’ll likely get a draft with lots of non-controversial material. Where’s the thought leadership in that?! Where’s the humanity?

Expressing ideas that people can push back against injects humanity into your thought leadership. Consider taking stands or suggesting actions in your thought leadership that people can disagree with. That’s another signal that a human is behind the thought leadership, injecting lived experience and personal opinions, not relying solely on AI to produce boring content.

Now, you might disagree with me on that, but that’s exactly the point. 😉

Fourth, apply the “Anybody else?” test to your content

Once you’ve drafted your thought leadership with AI assistance, a fourth way to make it sound like you is to apply the “Anybody else?” test.

After reading the draft, imagine a colleague’s or a competitor’s name and photo on the article, then ask yourself whether anyone would have a hard time believing it came from them, not you. 

Obviously, when attorneys in the same practice areas or serving the same industries talk about the same topic, there’s going to be overlap; it’s inevitable. But if your draft carried someone else’s byline, would other people think nothing of it? Is there so little of your personality, lived experience, and/or opinions in it that anyone could have written it? If so, you’ve got some work to do.

If your draft fails the “Anybody else?” test, I recommend you tinker with it until it has a bit more of a “you” flavor, even when covering topics your colleagues and competitors could and would cover in their own thought leadership.

Yes, you can use AI to draft content that sounds like you

If you want to use AI to help draft your thought leadership content, I won’t stop you. In fact, I encourage people to explore using AI as part of their writing process. 

But most people know by now that AI can easily whip up content that’s strikingly similar to what it produces for other people prompting it similarly. Your content is too important to your marketing and business development efforts, and to your overall personal brand, to be bland and sound like anyone else could have written it.

Your thought leadership should reflect your experience and personality, both of which drive why people hire you. The less it reflects both, the less effective it’ll be at persuading people to do so.