Use the Tools.
Keep Your Voice.

Transcript from the episode recording, with speaker names added and paragraphs formatted for readability. Automated transcription may contain errors; the audio is the original record.

Brad Hauck 00:14

Whether you're a speaker or an author, your voice is your greatest asset. Yet right now, AI tools can help you write faster, polish ideas, and even draft whole chapters in less than five minutes. That convenience sounds great until you stop and ask the harder question. At which point does the work stop sounding like you and start sounding like the machine? This episode is not about fear or hype, it's about responsibility. If you're using AI in your writing or your stage prep, you're shaping how your audience sees you, trusts you, and believes your stories. So today we're going to pull apart some of the ethical questions that most people skip, not to judge, but to raise your own standards.

If you're serious about protecting your reputation while still using these really powerful tools, this is the conversation for you. Good morning, Dixie.

Dixie Carlton 01:00

Good morning, Brad. I love the sound of this conversation.

Brad Hauck 01:04

Yeah, look, it's something that's coming to the fore. And I even notice it myself because I use AI quite a bit when I'm planning podcasts or videos, put my content in and then get it to give me some extra points. Like you do. But at what point are you actually putting forward its content, not necessarily your content? And is that a bad thing? Because sometimes because it's got so much knowledge, it can take what you've got and add a pile to it. So there's so many questions about where do you start and stop and where does it start and stop, I think.

Dixie Carlton 01:36

One of the things that you alluded me to a long time ago was the idea that you can create within ChatGBT your own specific library folders, if you like. And so I have created those with the intention that will not take it, will won't search the web for additional information. So for example, I'll set one up that might be Dixie Carlton Publishing and Marketing Expert. And then I will feed some of my books into that particular library or folder, and then have it only require it that it only goes into those areas to get information for my blogs, for example.

So that it is already checked and edited content, it's already my own content. And I've specifically asked the ChatGPT not to go outside of that. So I'm confident that when I create that, it does actually sound like me. And when I go back and humanize it further, I then check for things like how many bullet points actually look like they've come out of AI. Do I need to change that? Do I need to change the M dashes and the N dashes? And of course, I've done a blog recently about how AI did not invent bullet points and M-dashes.

Find them even in very old texts. But the constant overuse of those can be a dead giveaway that something has been created on AI. And so now people will look at that sort of thing and assume that just because there's a few M dashes in something, or ticks for bullet points, that it is 100% AI generated, which in fact it's not. So I think that there's an element of education there that we need to be applying to some of our work when we're putting forward our work and our pitch documents or our contracts need to specify whether or not AI is used and to what degree.

So I always explain this point to my clients that I will, if I create a folder for them, that it is ring-fenced with only the information that that pertains to them, and that therefore their information or what I put forward for them is not over influenced by ChatGPT. Or it's not being used to train chat GPT, which is another thing. And I find that helps a lot with the humanizing and the retaining of the voice. And sometimes it really does come down to how well you train or set up each of those folders.

Brad Hauck 03:59

Yeah, there's a big push these days, certainly in automation, for a thing called guardrails. And so there's actually like a block that you can use when you're building automations now called guardrail, and you set it. And the idea is that you can set these rules that it can't break, such as when their data comes out, it cannot contain phone numbers, or when the data comes out, it must have this randomized or whatever. So there is definitely a push down this route of isolating information. And you may or may not have noticed, but some of the tools that we use now, like Gen Spark, it now has a local-only mode.

So what it does is it will download a version of Lama or one of the other platforms that you choose, Gemini, to your computer. And it will work only on your computer. And so that controls what's being fed back. But also, as you said, using folders and doing your settings is absolutely essential when it comes to protecting the data. One thing I'd like to think about is what counts as your voice when AI is in the mix? How would you describe your natural style if someone removed all the AI from your processes, for example?

Dixie Carlton 05:12

Well, first sat down one day and when I was first learning how to use Chat GBC, and I used the deep dived into the prompt where I analyzed my own writing style and then create one of my folders to be more like my style. And we had some funny things happen with because at the beginning, because of my name being Dixie, it proposed all of these styles that were real down south, south and get your cowboy boots on. And it was really bizarre. And I thought, no, no, that's not me. And it was purely based on AI, thinking the name Dixie was Deep South America.

And it was like, hang on, that's not right. So I went in and I played around a lot more and got it to analyze specifically my style based on some of my books. And I found that a lot better as a means of saying, but this is my style. This is my voice. So now I have ChatGPT for my own stuff. If I ask it to write in my voice, it does actually sound exactly like me. I can't fault it. And I've got another two clients where we do this a lot when we compare notes sometimes.

And when I've gone to write to develop her blogs or work with her on her blogs, we have tapped into the ChatGPT system of her being very much her voice as used for writing this material, and she can't fault it. She says, Oh wow, she said, Have you already humanized that? And I've been working with her for a number of years, so it was really easy for me to know how to write like her anyway. And I'm a ghostwriter, right? So I know how to get into someone's head and how to pick up their style.

But the AI was incredibly switched on in some places where we couldn't fault what was just being produced. But again, that comes down to knowing what your style is already. Listening to someone. If I'm ghostwriting, for example, I will listen to the interviews, I will listen and I will pick up the inflections and how they speak, what kind of languaging they use. And I think if you become self-aware about how you do, that's going to help you protect your own style and your own IP.

Brad Hauck 07:20

Yeah. I think the biggest mistake most people make, in my opinion, is that they rush to use AI. In other words, you just go in and say, give me an article. Whereas if you look at what you did, and it's the same sort of thing that I do with using folders to separate your information out and to train it into a voice for that folder. You're actually taking time. So if you sit down and force yourself to create 10 pieces of content. Now, how you create that's up to you. You can do it verbally and then edit the transcription, or you can sit there and write, it's totally up to you.

You can take your books, but take the time to create a series of pieces of content and then give it the content and teach it from the content and get it to analyze the content and then get it to write something, a couple of paragraphs, nothing too long because you don't want to be spending hours, but a couple of paragraphs, and then go in and edit them to how you would write them. Now, sometimes I look at stuff that comes out of AI and I'm like, actually, that's better than how I would have put it.

And I might leave it there. But the point is that you want to then give that back and say, I have rewritten this. This is how I want it to sound. Learn from this and adjust a prompt. Whatever. So there is that element of preparation, training, testing, editing, improving before you get there. Now, once you get there, you're good. It'll never forget. And that's the key thing that most people forget. They see all the hassle they've got to go through. So they just go, I'll just write it. And I get that because we're always trying to produce more content all the time.

Everybody goes through that. But are you actually producing stuff that's you and will anyone sit down and actually read it all anyway?

Dixie Carlton 09:04

That's the thing. And this comes back down to also using good stories that AI can't create your stories. Only you can share your stories. Only you can write your stories. Because AI is not inside your head. You know what happened in that conversation between you and that person in the hallway at that conference. AI can't write that.

Brad Hauck 09:24

I beg to differ. So here we go. I was writing some content and I wanted some stories to go into that content. And the AI actually wrote the stories. And I'll tell you right now, they sounded like something that I would say. They were very real and believable, except that they just never happened. It actually blew me away because then I was torn, and this is where this ethical debate comes into it. Do I leave these stories in? Because I've actually experienced this a dozen times. It's just that I never told them that I told the AI that I had done this.

But these stories are very believable and they illustrate my point perfectly. And is that any different to using a fairy tale or some story to illustrate your point when you're speaking? Does it have to be real?

Dixie Carlton 10:17

He's a sideways look at that. Often I'm dealing with people wanting to write case studies or use examples of something that does happen in their business as an example, as a case study. But it may be that it's actually two case studies blended into one, or three or four versions of how this might happen. And you take it and you say, instead of it being this couple that did that, it's now a younger person, a student, instead of being an executive or an entrepreneur, it may be that it was a uh young mum, for example, I don't know.

But you change how it comes out so that instead of having this person and this person, you actually say, Bob, who was 35 years old, did this thing, and this was how much better his car operated from based on these particular things that we talked about. And so you can blend good case studies and write them as one.

Brad Hauck 11:11

But is that a lie? If someone came up to you and said, Did Bob do that? you'd have to say, No, Bob did not. It actually was a couple. And so there was two people that dynamically were involved in this situation.

Dixie Carlton 11:23

But then what I also say to my clients when we're talking about doing this is if someone says to you, Did that really happen, did Bob really drive that car in that way and get this massive news story as a result of, I can say, or I'll say to my clients, by all means, have the reality of that behind you to make sure that if you need to be checked on that, someone can go and talk to those people. Because maybe you're protecting their anonymity because of whatever reason. Sometimes we're dealing with lawyers and medical professionals, and you can't actually be too obvious about who that person is.

But if you've got someone who comes to you and says, I think you've just described my sister who lives in Cape Town, and I'm offended that you wrote that about her without her permission, you can actually go back and say, no, if you really need to, you can go and talk to Bob the builder and Thomas the tank engine and actually figure out this is how that story got blended. So you have a very strong level of ethics required to do that. And I think this really does come down to an integrity issue. If you are going to write something that really has been created fully as a story, as a fantasy by AI, then you have to say that.

But you could say this is to demonstrate, and this is based on two or three typical scenarios in my business.

Brad Hauck 12:49

Yeah. You can do the old animated cartoon type thinking. Bob and Mary walked up the hill. So you can do that, that's fine. And AI will write those stories for you to illustrate the point that you're trying to make. Or you can give it the facts and get it to write a story around the facts that doesn't have the original people in it. So you can do those sorts of things easily enough. But it is an interesting one because I've seen some really great stories get thrown out of AI, and I'm like, oh, I'd really like to use that.

That's really good.

Dixie Carlton 13:18

Well, you just have to be really honest because the problem that you have is that if you go too far down the path of using that, then all of the real stories become less believable. And if people think, oh, he writes a really great story, and that sounds really interesting, but then they meet you and you can't actually tell a real story anymore. It's like using too much of Surrey to read the maps. After a while, your actual own brain stops having that storytelling ability that you used to have.

Brad Hauck 13:50

And I think one of the things that I like to do actually is if I do come across a story that I think works really well, what I will do is send I off to do research of where that happened in the real world. And it'll often come back with, oh, this company, Peppers, did this or ROCQ did that. And it's the same story, but it's actually a real world example of where that happened. And then you can use the story, but use the real world example. So you can use the AI to give you ideas, but then send it off to actually come back with something that's real.

So don't ever get stuck on that stuff. But if you're ethically thinking about those and you go, I like it, but it's not me, find a real world example of it. There's billions of them online.

Dixie Carlton 14:34

You're so right. And then the other thing too is if you're even if you're speaking on stage and you're telling a great story and someone puts their hand up in the audience and says, Yeah, don't believe it. That sounds too far-fetched, not possible. Then you actually have to be able to say, I actually know the person who was involved, and this was the additional piece of information that might make it more credible. I'll give you a great example that's just come up, in fact. I'm getting a lot of really good, interesting history stories coming up on my Facebook feed, and I'm loving them because I'm really interested in how those stories relate to the settling of the Wild Western America, for example.

And I have wondered a few times whether these stories are in fact all real, because some of them just too.

Brad Hauck 15:20

I know exactly what you're talking about.

Dixie Carlton 15:22

The two small children that raise themselves on a farm in the deep in the wild south all through a winter and buried their dad in the snow and all that sort of thing. Those are very inspirational, but I don't necessarily believe that they are all true. This morning I woke up and there was a story that was of that nature with a photo on something that happened at the end of World War II, where a plane took off from Vietnam, had a father and his wife and five children on board this teeny tiny little plane, and it managed to land on an aircraft carrier that the there's a whole lot of story behind it.

Anyway, the story was posted or shared by someone that you and I both know who we knew that he was, in fact, in the Navy in 1975, and he had a photo of himself that I could recognize with that particular aeroplane. It's quite a famous story. I'm not going to share all the just spoiled if you don't have to go and look for it. But that made me rethink the fact that I know someone who was able to verify that this amazing, far-fetched sounding story was real.

Brad Hauck 16:30

Yeah, there's a lot of that stuff. And in fact, as you said, there's huge feeds of that stuff. I've had a pile of it coming up on my Facebook feed late lately. And I've questioned it sometimes and gone and checked because I thought the story was great, but was it expanded upon a bit by AI? Because you could tell it was AI written. It was accurate, it was just filled in. Yeah, there's definitely that.

Dixie Carlton 16:52

As writers or as storytellers, and if we work on stage, we have to be storytellers. If we write books about non-fiction topics, we still have to be storytellers. And if we dilute our ability to write and share and speak good stories, then we're not doing ourselves any good service long term. We have to still be writing our own content and being good storytellers so that if we are cornered or asked directly, was that for you for real? Did you write that? We have to be able to say yes.

Brad Hauck 17:24

And if we Yeah, absolutely.

Dixie Carlton 17:26

Yeah, and if we lean into I'm gonna use AI to tickle this up 10% or 90%, then we at least do it to our audiences to be honest about that.

Brad Hauck 17:36

Yeah, and look, quite frankly, the better the story is, the less people are gonna believe you it's real anyway, half the time. It's not like no, that couldn't have happened. Like people are pretty skeptical. So yeah, look, there's so much to do. And I think another aspect, are you transparent about this stuff? What level of openness do you say to people? Actually, I use AI to write this, or I use AI in my planning. It does come to a point where you think, how do I put this? For example, I do use a lot of AI to do tasks.

I've trained the AI to do it my way. But does that mean it's no good now because I can press a button and have all that work that used to take me weeks come out in an hour? And do I tell my clients that I use AI? Yes, I do. But and it's quite obvious by the volume of work sometimes that you couldn't possibly do it without AI. That's the other thing. Is that ethically bad? Do I should I feel bad about charging for that because I didn't do the work? I don't feel bad because you couldn't have done it on your own, number one.

Number two, I spent years learning what to ask and how to get it out. Number three, I developed the system. And number four, I took the time to actually create the content for you. But there will come a time when you're asked, people will say to you, This is obviously AI content. I'm like, yes. Why should I pay you $10,000 for that?

Dixie Carlton 18:59

And you know, I think we also have to be aware that the readers and the audience members are starting to become so discerning that they can pick up very quickly and very easily. Especially our editors, like my editor can pick up on something that's AI generated feeling really quickly. I can usually up to a point as well now, but she's even better at it. You've got to know that if you are writing that way, that you are going to become known as either someone who is someone that uses AI let's talk percentages, 10%, 20%, 50%, 90% of the time.

How do you want to be perceived? How important is it for you? Like for me, it's important for me to have my stories be believable because I write fiction as well as non-fiction for myself. So I need to be known as a storyteller. I do not want to have my reputation tarnished by having people think that I create any of my stories or any of my written content using AI. I will use AI as an assistant to do the things that I know that would otherwise take me longer. Regularly for I'll take in bio, for example, and if I've got enough information, I might feed it a couple of extra pieces and then I'll put it into a chat GPT and regenerate a bio.

But it's still very much a bio that sounds believable because it's based on an existing bio that's already been written and approved by the person who it's about.

Brad Hauck 20:25

Yeah, I think like I create a lot of my content like this podcasts, videos, things like that, rather than necessarily writing my content. I start from audio or video because it guarantees it's my voice. Yes, I might have an outline done by AI that I've fed at what I want to cover. It gives me some extra bullet points, and then I go through. But you'll see that I switch back and forth between what I've listed and what I'm thinking. And so I'm getting this mixture of I'm being guided into the content, then I'll discuss it, then I'll come out and be guided, move back forth.

So I think you know, looking at how you create content may affect, as we go long term, how authentic you can keep it. Because for me, writing is hard work. For you, writing's easy. But for me, speaking is easy. I've got to look at, like I could sit down and write, but man, that's hard work.

Dixie Carlton 21:16

This is one of the things that comes back to how we want to be perceived. So, for example, with podcasting, which is the new black, we all know that podcasting is the greatest best thing for getting your marketing accelerated. You and I created this as a podcast, as a conversation, because we wanted to have real conversations and authentic sounding hearing of exploring the topics that are important to experts and authors and speakers. These days, and I met someone at the Book Thinkers conference recently where I did question whether or not they're doing the amount of podcasting that this person does, whether most of the content is read content because it's AI generated first.

A number of people I know are putting out five, ten, fifteen minute podcasts. And it's just them. And they're just speaking. And you can tell when you start to listen whether or not they're reading something versus just something. Sitting there and shooting the breeze and having a conversation with themselves or potentially with their listeners. So when you start to hear those people and you can hear that there's that slight subtlety of the reading the content, it's a teleprompter thing. You have to wonder whether that teleprompter has also been already written, the content already written by AI.

Brad Hauck 22:37

Yeah, I think look, man, I do use it in that way, but I think that it also keeps you on track and it's like having someone in the studio guiding you. Yes. While yes, you are reading some things because you want to make sure you get your points across and then you discuss them and talk about them. Yeah. I get what you're saying. I don't necessarily agree that it's actually an issue, but I think the key thing is that we've always done it that way. We've always had our list of bullet points we're going to talk about.

That's what we're doing. And it's very hard when you're doing a solo podcast like the Profitable Speaking podcast that I do to have a conversation with yourself. So you do those points to kick your head and get your brain working a bit. So definitely one of those things.

Dixie Carlton 23:20

But my point is if we are only listening to podcasts that sound like they have been recorded in a studio based on written content, we listen to it differently than if we're listening to two people having a conversation. If we just use bullet points and we go off script sometimes with those bullet points, we sound different. And so if we are doing the solo style of podcasting with just one person, you have to be willing to not only rely on what you've written and read it out. Because listeners will become far more discerning than that.

And that's where I think using too much AI is going to damage the long-term viability of some of that kind of content.

Brad Hauck 24:06

Yeah, I think one of the things that I've done to actually overcome this, and look, it is an ethical debate, there's no doubt about it. How much is too much? We watch documentaries on TV, they read every word, and we enjoy them. And then we watch back shows and they talk about things. So they are different necessarily genres of podcasts. But one of the things that I do like to do is I actually have my little main point, but then underneath that I have bullet points, which are just one or two words. So I don't have long readable sentences, and that forces me to actually come up with content that's that keeps me on track, which is what the goal is in the end.

So there's so much to look at here with AI. And I think some of it comes down to what you'll accept. And the other side is does it actually resonate with your audience? We always come back to that. So if you're doing it and no one's listening, then maybe you're not resonating with your audience. Maybe it's actually not working for you. Maybe you need to look at it and say, Am I using AI too much? Is it no longer me? Am I really doing a discourtesy to my audience?

Dixie Carlton 25:09

Yeah, no, I think we've raised some really interesting topics within this one.

Brad Hauck 25:15

Absolutely. What a conversation. So AI can move faster, write cleaner, and polish your message, but none of that matters if people stop trusting the person behind the microphone or the person who's writing the page. The challenge now is to use AI without handing over your personal voice and your credibility. You need to ask yourself the tough questions, some of the ones we've looked at today. You need to set rules for yourself and processes in place. And you need to build workflows that prove the value that you bring to your audience. Thanks for tuning in to the experts journey.

If you picked up some useful tips today, take a moment to subscribe and share this episode with a fellow speaker or expert in your network. It really helps us help people. And don't forget to check out our books on Amazon, Dixie Start with the Draft, How to Easily Plan and Write a Nonfiction Book, and a Dynamic New Conference in a Book Series. The first one being Connect Leads Succeed. And my book, Micro Course Profits, Build a Seven Figure Income with Bite Sized Micro Learning Courses. Both are packed with practical tools to help you grow your business faster.

Catch you next time.