Ramona Shaw
If you've ever read an email from someone on your team and thought, "Hmm, this sounds like ChatGPT," you probably took it a little less seriously right in that moment. And your team is doing the exact same thing with you. So today we're going to talk about how not to sound like ChatGPT as a leader. And I don't just mean the em dashes, and I also don't mean getting rid of ChatGPT at all. I mean how you navigate and curb the cost and the downside of not sounding authentic because something else actually scripted your message.
My name is Ramona Shaw, and this is The Manager Track Podcast. So here's what I'm seeing more and more. Managers are using ChatGPT, and by the way, I'm going to use that as a synonym for all the other products like it. It could be Claude, Gemini, Copilot, you name it.
They use it to figure out how to give feedback. That's a big one. They ask it for a script for a difficult conversation. They have it write their updates: their team meeting updates, their status updates, the end-of-week update, their performance reviews. Yes, their self-reviews and their feedback for other people, even their meeting agendas.
And on its own, that's not a problem. It actually sounds like a great productivity gain. Something else is doing a lot of the heavy lifting for you. But the problem is when we use it without understanding the psychology behind what we're saying, and without actually having the mindset that would lead us to say it in the first place.
So we're kind of taking the whole thing the wrong way around. We're starting with saying something that should actually come at the end of a chain reaction: from thinking a certain way, to then doing something a certain way. Now suddenly we're doing something a certain way without understanding why, and without having figured out the mindset part at all.
So we're going to cover why this comes across as off even when the script itself is good, what it's costing you, where it goes wrong, and what to do about it, so you can actually use ChatGPT and the like to reap the benefits of it, to be more productive, more effective, and better with your communication, without the costs that come when it sounds off.
So let's start here. The way we think directly influences how we show up. When you genuinely believe something, let's say you believe your employee is capable of a lot more and you really want them to get there, that shows up in your tone, in the words you pick, in the questions you ask, in what you notice.
It's consistent because it comes from the inside out. Now, when we do this backwards, we show up a particular way with words and a structure someone else handed us, but our thinking doesn't match, and that mismatch is what people pick up on. So the script might actually be a good script. On paper, it would work.
But people spot pretty quickly that it's not genuine to the person saying it. It just doesn't fit you. And here's what's new with AI, and why this is different. Before AI, people would pick up on that and maybe have a gut feeling. Something feels a bit off. Maybe you were being coached. Maybe you read a new book and you're trying something new. Okay, that's part of behavioral change. It just doesn't feel authentic yet. It's part of learning and growing. We're adapting and changing.
But now they have an explanation ready to go. It's right there on the screen: the reason you sound weird and inauthentic is that you were lazy and you used ChatGPT. You used another interface that told you what to do without you really doing the work. If someone is genuinely interested in behavioral change, they're getting coached or learning a new skill, and they're trying to say something different than they used to because they know it would be more effective, you may realize, "Yeah, it sounds a bit inauthentic, but I appreciate your effort to grow and learn."
But when someone uses ChatGPT, our assumption is that they're not really trying to learn and grow. They're trying to go fast and easy. And so you don't get the benefit of the doubt. It's a net negative. And what makes this even harder is knowing that people will land on that conclusion even if you didn't use ChatGPT at all.
So when something starts to feel inauthentic or scripted, people think of AI. There's research that backs this up. In experiments at Cornell, researchers put people into chats and offered AI-suggested replies. When people assumed their chat partner was using AI-suggested replies, they saw that partner as less cooperative and felt less connected to them, whether or not AI was actually involved.
They didn't know for sure, and not all of those chat partners actually had AI-suggested replies. The suspicion alone did the damage. So now let's be really clear on what this costs you. The first cost is a loss of authenticity. The second is a loss of connection. And the third, which may surprise you but shouldn't, is a loss of trust in your capability and your competence.
The Stanford Social Media Lab and BetterUp Labs published research in Harvard Business Review on what they call workslop. This is AI-generated work that looks polished but doesn't have the substance to actually move the task forward. About 40% of the desk workers they surveyed had received it in the past month, which is pretty significant.
About half said they saw the person who sent it as less creative, less capable, and less reliable. And 42% saw them as less trustworthy, which is pretty big. So people aren't just judging whether the document itself is good. They're judging you, and your use of AI.
Now, when it comes to leadership, where creativity, competence, and trust matter significantly, this is something worth paying attention to. If people don't trust you, or they don't feel connected to you because you're using AI, or it seems like you're using AI because it just doesn't sound or seem like you, then the relationship starts to become transactional.
You say the thing, they do the thing. But over time, that hits motivation, engagement, your long-term relationship with people, and even creativity. People stop bringing you their real thinking when they don't believe you're actually curious and interested in working through the problem yourself, and when they think you rely too much on AI tools.
Now let me walk you through where I see this happen all the time. And let's be honest, where have you maybe overdone it? We're all trying to calibrate. How much can I use it? Where can I not use it? Where is it hurting me? Where is it helping me?
Many of us who have started to use AI in our workflows have probably overstepped that fine balance at least once, if not multiple times. So it's all about becoming aware and calibrating. We see it in feedback and performance reviews. This is now the season for a lot of this.
Review feedback that sounds balanced and professional could have been written by anyone. I often get 360 feedback, and I can immediately tell which comments were written by AI. Now, some people use AI in a way I can never tell, but there are some very obvious cases, and the obvious cases are where the trust gets lost.
In those examples, there's usually a strength, then a growth area, then a nicely phrased goal for what they could do next. But when the reader notices that AI wrote it, they hear, between the lines, that you didn't seem invested enough to write this yourself, that you didn't pay enough attention to come up with your own sentences and your own judgment.
Now, that may not be true at all. Maybe it was eight o'clock at night, you had to knock out 10 of those reviews, and the only way to meet the deadline was to put your bullet points into ChatGPT and take what it spit out. It's still your own observations. But the problem isn't what you did, your intention, or what you put into it.
The problem is always what the other person perceives. And in that moment, they perceive that you don't really have the same skin in the game they thought or wanted. The other place it often comes up is agendas, reports, and proposals.
These are the ones that just feel outsourced. Everything is in there. It's organized, it has headings, a great executive summary. But when it comes to the substance, and especially the person's conviction and their unique angle, it can feel empty.
There isn't really a strategy that someone owns or could defend if someone pushed back. And again, I'm not saying that just because you use AI, you don't have a good proposal. The problem is when you don't have a good proposal and you patch it together because AI helps you. Then it turns into a problem.
Any time we're looking to elevate our communication, it's never to compensate for a lack of knowledge or competence. Communication sits on top of your knowledge and competence. And the same is true of using AI to improve communication, which gives us a bigger lift, or at least a quicker lift, than any training, framework, or communication skill ever has before.
So if you have shaky, missing pieces in your competence or your knowledge, and you're using AI, it seems like it's filling all the gaps, and it looks great. Except for the moment you present it, or someone reads it, and they start to think, "Wait, that doesn't make sense. This doesn't hold up." Now you have a problem. You're losing credibility really fast.
The same is true for conversations. Not just performance reviews, meeting agendas, and proposals, but also the feedback conversation: "Hey, I want to tell you something that's not working well. Let's talk about it." Coaching conversations, difficult conversations that are basically scripted.
Let's say a client of mine, I'll call her Chen, asks ChatGPT for a script to address a performance issue. This one is fictional, but I see it all the time. Her employee says something she didn't expect after she delivers her script, and now Chen is off the page. All of those lines seemed so great, but in the moment she doesn't know how to adapt, because she didn't understand the reason for the script in the first place. And the employee can immediately feel it. They can feel when it isn't genuine.
The same is true for emails, and this one is sneaky, because often there was real thought behind the email. You knew what you wanted to say, then you ran it through AI to make it sound good, and it came back whitewashed. It took the gravitas out of your message.
Sometimes even the emotional urgency, the kind of seriousness you wanted to bring across, all whitewashed. It sounds great, but it also sounds like everyone else's email. And if it's obvious right away that it's AI-generated, people might not take it as seriously as it deserves. So you might be putting in all the work, and then it lands flat.
Now, I want to quickly pause here. If this is resonating and you're thinking about how to be an effective leader, how to build trust and competence, understand why feedback is given a certain way, how to coach without a script (because coaching can't be scripted in the first place), how to motivate employees, and how to navigate conflict and difficult conversations, those are all important skills. And if you realize, "Hey, I haven't quite figured this out yet. I don't know what these frameworks are or how to develop these skills. I don't understand the psychology behind it," then I have the exact right program for you.
I'd like to tell you about the Leadership Accelerator. This is a 12-week program that we run for new and early-stage managers, people who are good at their job and now need to become good at leading. We go deep on exactly these things: giving feedback, performance conversations, coaching, conflict, how to delegate well, and the list goes on.
That way you don't need to rely on ChatGPT for scripts without understanding why they work, or how to pivot in the moment when someone goes off-script, because you understand the foundation. You have the skills and the competence that AI can't compensate for.
The Leadership Accelerator includes on-demand training, so it fits your schedule and stays flexible, plus live calls and one-on-one support, so what you learn actually sticks and you apply it week over week over the course of 90 days. This is a proven and highly effective program that we've run for years now, with hundreds of managers making leaps in their leadership effectiveness and their confidence. To learn more, check out the link in the description. And with that, back to the episode.
So let's talk about what to do about this. And no, I'm not going to tell you to stop using AI. I think AI is an integral part of our work. But in order to avoid the cost of not being seen as collaborative, competent, or trustworthy, we have to start with the leadership philosophy.
If that's not in place, you're going to be inconsistent, and you're going to come across as inauthentic. And honestly, you might be. You're probably not being fake on purpose; that's not what I'm saying here. But you might be reacting to moments like, "Never seen this before. Where is that coming from? Hard decision here, new decision there."
So you're shooting from the hip, responding to whatever comes up, and then asking ChatGPT what to do, instead of being grounded in what you stand for as a leader and what your leadership principles are, and being able to answer questions such as: What do you understand about people? Why do you give feedback in the first place? What do you value in a feedback conversation? What do you believe someone is owed when they fall short?
If you can't answer those and you're just reading a script, it's likely going to backfire. This is really about critical thinking. You need to understand why you're saying what you're saying. If the thinking isn't done, no script is going to cover for it.
So before you open any tool, ask yourself: What do I actually think here? What outcome do I want for this person? What matters to me here? Which leadership philosophy, principles, and values do I demonstrate with this action or this message? That is where it all starts.
Now, this actually goes beyond AI. A lot of leadership development happens through micro-training: micro videos, LinkedIn Learning, and maybe even quick tips from social media. But those rarely explain why we do what we do. So the skill stays superficial, and superficial leadership skills will carry you for a few months, but not for years.
With AI, information is right at our fingertips, so knowledge on its own is becoming a lot less valuable. What's going to set people apart is wisdom. I like the Cambridge Dictionary definition here: wisdom is the ability to use your knowledge and experience to make good decisions and judgments.
Having the information is the easy part now. Knowing how to apply it with this person, in this situation, is the part AI can't do for you. You have to actually understand it. It can't do the understanding for you.
And if I put a bit of a Stoic spin on top of that, the Stoics described wisdom as knowing what is good, what is bad, and what is neither. A big part of that is understanding what's within your control and what isn't, and then making sound moral choices from there.
In a feedback conversation, you can't control how someone receives what you say, but you can 100% control whether what you say is true, fair, and actually yours. And you understand the reason why these frameworks work, so your message lands.
Let's get a little more practical. There are four things I would do. First, think before you prompt. Write down what you actually saw, what you want, and what you believe about this person before you open any tool. If you want to prepare for a feedback conversation or write something for a performance review, write it all down before using AI.
That is your stance. That's the foundation. AI can help you sharpen it, but it shouldn't be the one creating it. You can even write it down and ask, "What is confusing? What is unclear here?" Instead of asking it to rewrite, it's a lot better to prompt it to challenge what you wrote, so you see what's missing, what's confusing, and how you could explain it better.
Second, teach your AI your language. AI will whitewash what you're trying to say if you let it. It will all sound the same if you let it. So give it your writing. Pull together emails and reports you wrote before you started using AI, and use them to set up a writing style it has to follow.
There are a lot of tools out there to humanize AI, but that humanizing tool is also just an average of all people. Create your own humanizer, based on your language and your word choices, not those of AI. That alone will make your drafts sound a bit less fancy, but that's the point. They'll sound more like you and less polished.
Third, get rid of the obvious AI signals. Even if you don't use AI, those signals immediately give people a bias about where the message is coming from: the em dashes in every other sentence, the phrases nobody actually says out loud because they sound so odd, the word choices you've never used before. Suddenly you realize, "Why am I saying that?"
It's because AI uses it. I've noticed that myself too. It's really easy to fall into, and suddenly you start to question why you're doing it. Once you have this sorted out, you can run your draft through your own humanizer so it actually sounds like you and doesn't get dismissed at first sight.
And fourth, read it back and check. Is my philosophy in here? My stance, my values? Do I understand why I'm saying each part? If I get pushback, if someone questions me, can I explain myself and this decision? If you need the script in front of you to say it, it's probably not yours.
Now, I do want to make an exception here. If you're learning new ways of saying things because you know they're more effective, scripts are actually helpful. I'm not saying never use scripts. But if a script feels so weird that you would never say it, consider that a signal. If it's that far off for you, it's likely going to come across as weird to other people too, and you'd have to practice and rehearse it a lot.
Understand the main message of the script, because that's what the script is useful for, and then edit it until it sounds like something you would actually say, something more authentic to you.
So here's what I hope you take away from this episode. Your team isn't looking for perfectly phrased emails from you. They're looking for evidence that you thought about it, that you care about them, that you mean it, that you stand behind it, that you have values, that you're consistent, reliable, and trustworthy. They check your communication against those filters.
Are you consistent? Are you fair? Can I trust you? Is there integrity in what you're doing? Are you making good judgment calls, or at least, even if I don't like them, can I understand where you're coming from based on your values and your perspective? That is what they care about. So if AI whitewashes all of this and takes it away, people are going to look at it with a bias. That's the risk, and it's going to hurt not just your communication but your leadership effectiveness.
If this was useful, please subscribe and share it with a colleague, friend, or team member who would find value in it as well. I'll add a few related episodes to the show notes, so check those out. And remember, the tool can find you better words, but it can't care.
Starting with last week's episode, many of our episodes now come with a downloadable guide, so check the show notes to grab the cliff notes with the main takeaways from this episode and keep them as a reference. With that said, I'll see you next week with another episode of The Manager Track Podcast. Bye for now.
If you enjoyed this episode, check out two other resources to help you become a leader people love to work with. The first is a free masterclass on how to successfully lead as a new manager. Check it out at archova.org/masterclass. The second resource is my bestselling book, The Confident and Competent New Manager: How to Quickly Rise to Success in Your First Leadership Role. Check it out at archova.org/books, or head over to Amazon and grab your copy there.
