Hello, and welcome to the Nonfiction Authors Podcast. I'm Carla King, your host, and before we start, I'd like to invite you to go to the Freebies tab at nonfictionauthorsassociation.com to check out our free reports. We developed these reports to help you figure out things like ISBNs, distribution, optimizing book sales on Amazon, generating book reviews, growing your email list, and we provide checklists on things like publishing and book launches. Now, stay tuned for this week's guest. Hello and welcome to the Nonfiction Authors Podcast. Today we're talking with Lisa Thee about the ethical use of AI for writers. But first, this podcast is brought to you by the Nonfiction Authors Association, a supportive community where writers connect, exchange ideas, and learn how to write, publish, promote, and profit with nonfiction books. Subscribe to this podcast on your favorite podcast app and visit our website to find transcripts, show notes, and links to all of our episodes. Also explore our membership options and download free reports, search the archives, and get answers to all of your writing and publishing questions. Now, I'd love to introduce our guest, Lisa Thee is a thought leader and expert in the fields of artificial intelligence, career transformation, and ethical technology. As the CEO and co founder of Minor Guard, an AI software company dedicated to online safety, Lisa combines her technical expertise with a commitment to ethical AI practices. She's the author of a new book called Go! Reboot Your Career in 90 Days and co author of Demystifying Artificial Intelligence for the Enterprise. She hosts the Navigating Forward podcast and her TED talk is titled Bringing Light to Dark Places Online, Disrupting Human Trafficking Using AI. You can find more about Lisa and her passion for harnessing the potential of AI to create positive change. That's Lisa T H E E. com. And Lisa, welcome to the podcast.
Lisa Thee:Carla, I've been looking forward to talking with you for quite a while. It's a pleasure to be here.
Carla King:Thank you. Artificial intelligence is such a hot topic now. We're a little bit confused about it. There are an incredible number of tools. Some authors and writers are thinking, this is taking our jobs. AI is writing books, but really what I want to do today is concentrate on how authors can use it for good and ethically and pick and choose through all of the tools. Not all of them, cause there's so many now we can't even list them anymore. And just use them with confidence. Your whole job lately, or for a long time has been about ethics. First I just want to say that I keep telling people, AI is inevitable. The Pandora's box is open. We can't put the genie back in the bottle. But everybody's trying to stop it, let's stop development. Do you agree that is not a helpful, how is the better question? How can we use it?
Lisa Thee:Yeah, I'm happy to. I think you mentioned this in the intro as well. There's a lot of trepidation about job loss as a result of some of these new capabilities coming in and the way that I try to ground myself and all of this waves of information coming at us is this feels similar to me like when personal computers came into the workplace. We all of a sudden had to upscale and it felt a little bit intimidating. Some of us were at different stages and so it was more natural to be in that learning process. Myself, I happened to be in college at that time. So picking up new things felt natural because we were in that learner's mindset, but there were people that were on the ending season of their career. And it probably felt a bit overwhelming to try to have to relearn how they do things. So first and foremost, I like to ground people. AI is not going to replace you at most jobs, but somebody that knows how to leverage AI may. And that's where you need to be focused. AI is a great rough draft generator. It's a great assistant. It is not often at this stage of things capable of replacing what a human can do. It's really good at those repetitive, tedious, pattern based tasks that are informed by what's happened, what's worked before. So I like to ground people in, yes, it is inevitable. The genie's not going back in the bottle. Just like you probably are not very employable if you don't know how to use the internet. But there's going to be levels of how in depth you need to be able to access it. So I really encourage people to think through how to just get more comfortable with it. You don't have to rely on it. It doesn't have to be in your everyday. To be perfectly frank with you, I have been building software and AI since 2015 and deploying it. I didn't use ChatGPT until March of this year. So we're all on this journey together. And I think that it's okay to be open about a little bit of discomfort.
Carla King:Thank you for that. And you made an important point there that, AI is not replacing us but we need to leverage it. We have heard a lot about whole books being written by AI. There have been some pro high profile authors who have been spoofed, with AI. But the tools, the publishing tools and companies are getting smart at identifying those, Amazon and IngramSpark.
Lisa Thee:I think that there are some tells with AI that you can tell that it's not quite right. It's not quite written by a human. The area that I'm more familiar with in terms of being able to augment is in the areas of video. So I think we've seen a lot happening with deep fakes where once AI is trained on someone's likeness, it can replicate with some degree of convincing what that person has done and make you convinced maybe they're saying something they've never said before, for example by using some of those tools. So I think the difference in terms of, it being a video or writing is fairly negligible once there's a lot of data collected about us and authors often are in the public domain. It does make you more vulnerable to spoofing because the computers can train on your specific voice. So I think that as fast as we're going to be developing tools to help us to accomplish things, we're also going to be developing tools to identify when people have misapplied them. It's not going to be a perfect science, but I think one of the places that I take a little bit of inspiration from maybe the example that we're seeing in the music industry where artists like Grimes, for example, are licensing their likeness for use in an AI model. I think that's going to be the way forward in that there's a way to get attribution for your unique voice and be getting some revenue from it that way by embracing the trend that's happening. I think also, going through a publishing process myself, there are a lot of checks and balances that happen to make sure that your words are your words, and it's not copyright infringement and everything is properly sourced and you have the right to use it. And so I think that inherent in the process, they're going to be catching a lot of these things. But I think that maybe people that are trying to leverage these new technologies are testing the boundaries like a child would to see if they can get away with it. But I can guarantee for somebody that recently has gone, from the beginning to the end of the process with a publisher, I think you're going to be in safe hands. On the self publishing side if anybody knows how to leverage AI optimally, it's going to be large internet service providers like Amazon. They've been optimizing AI for commercial applications for a lot longer than the rest of us trying to trick them. And so I have a lot of confidence they're going to be able to identify when something is not quite right.
Carla King:It's interesting. I've been talking to principals in IngramSpark, for instance, and Amazon and, they're using AI to check for AI and it's disturbing in a way, but it's also confidence boosting because the AI can very quickly go through the words and match them for plagiarism, for instance.
Lisa Thee:Yeah, and there's a great precedence there actually in cybersecurity. It's very often that you see AI deployed against AI to be doing threat detection and analysis is actually quite effective in those cases because it is able to process so much more data than humans can really hold all at one time. So any little tell is it's looking for pattern matching where they can identify that it wasn't created by
Carla King:And you work in the enterprise so I'm sure that you build a lot of those systems or know how they're built and are very in touch with how these large organizations are using it ethically and to identify issues with what they're putting out into the world.
Lisa Thee:And copyright infringement has been something that the industry has had to be identifying for eons. So I think in some new areas where AI is being applied, I think that's the big shift is it's more democratized. Now, you don't need to work for a large tech company that has been investing in AI, for over a decade that probably processes transactions for retailers and has, just huge volumes of data. There's probably 10 companies in the world that meet that requirement. So if you didn't work in one of those, all of this feels very fast and very new. But what I can tell you is the reason that it stayed concentrated in those areas is it takes a lot of money to build models. And it takes a lot of money to hire the people that know how to do it. And this is technology that actually has existed for quite a long time. None of this is new for people that work in this field. What's new is the awareness, the access, and then a democratization of the technology and, that anything new in platforms, it can be applied for good purposes. It can also be applied for bad purposes. The technology often is neutral. It's how the humans choose to use it that can become problematic.
Carla King:Absolutely. Thank you. And when we're self publishing, or if you're a small press or publishing house you can use AI for great things. You can do keyword research and, copywriting, things like that. What else can you think of?
Lisa Thee:What I really enjoy as a writer that takes so much pressure off my shoulders is actually on the marketing side of the writing process. So I think a lot of writers maybe meet the model of they love to be in control of their words and have the autonomy to be with their computers and expressing themselves freely. But then once you get past that publishing stage and you actually have to make the world aware of what you're doing, it's incredibly uncomfortable often to be self promoting. Maybe not what you anticipate when you go into a writer's lifestyle. So I love it for that rough draft generator when I have to do a social post, for instance, to make somebody aware of something that I'm writing about, or when I need to maybe do a newsletter to be connecting with my audience. It gives me so much more of a head start when I'm Feeling some writer's block about how to communicate a message that I spent two years creating, but now I need to get down to a 30 second blip to make people interested in it. I always deem myself to be a better editor, maybe, than an original writer, and so having something to go by, to me it almost feels like writing prompts from a publisher, in a lot of cases. Then I find myself doing a lot of editing because I can read it real quick and go that's not right. And that's not right. And that's not right. And then it's just changing those couple things versus being like, where do I even start? I think ChatGPT is great for that. I agree on that tool as something to be used. Because the information that I'm putting out there is already something I want the public to be aware of. But I think it can be really dicey as a writer to consider using things like that, because what you need to recognize is when a tool is free, that means that oftentimes you are giving your information away. So if you're writing something bespoke and unique that you haven't published yet, be aware that you are giving that information away or your ideas or all of those creative juices, so to speak to a company. And so we saw this with people that were using it for like press releases at large companies, maybe it was a few weeks before earnings and they were trying to draft that. Or perhaps there was an engineer that was looking for some information and they put. Some intellectual property in there. We saw that with Samsung. There was a big leak that happened as a result of one of their employees making that choice. So from an ethics perspective, it's really critical to realize when a tool is free, you're giving permission to use your information. So what I prefer on the marketing side is using a private tool. It's run by a woman executive, which I really enjoy in the startup world, seeing women thrive because they receive a lot less funding in that domain and it's called Lately. And what lately does for me is it takes all my long form content and creates an automated way to be getting that into digestible snippets using AI. So it writes the copy for me and it edits down my videos into smaller sections that reflect the copy in order to create an automated tool to publish a month's worth of content calendar. And it gives me. A lot of my time back because it does all of that editing and all of that rough draft writing and then I can just automate the publishing through my standard social channels. So I really love lately. because it allows me to keep my intellectual property my own, but it allows me to leverage the automation. That continues to learn on my post what performs and what doesn't and keeps tuning in my voice.
Carla King:That's interesting. I hadn't heard of that tool. So yay. But now you said video content, some somewhere in there, how does that fit
Lisa Thee:in? So I think a lot of A. I. There's kind of two domains of A. I. That have been dominant, and they tend to go back and forth into what is in fashion. So when I started working in A. I. In the 2016 17, it was all about image recognition. So using A. I. To be able to identify things in images and videos on being able to accomplish whatever task you're trying to do by matching those. In my instance, we were doing that with images of currently missing children from the National Center for Missing and Exploited Children with escort ad advertisements to be able to identify that this escort is a known missing child. This can't be a voluntary activity because children can't agree to participate in that kind of work. And so it was bumped to the top of the queue for detectives to be able to identify and recover that child and get them the services that they need to be rescued from human trafficking because children can't consent. So that's an example of how images are used today. What you hear a lot about is natural language processing which is that route of the ChatGPTs with generative large language models and what's changed between the days is a lot of it depends on the cost of compute and being able to do it more cost effectively to solve. Large challenges. In terms of video manipulation. I always go back to that for one reason and one reason alone, which is we've all been raised to see it, to believe it. What I'm most concerned about is we're entering a post truth era where it's one thing if somebody writes something about you, you can fact check it and go, Hey, this isn't right. But when you have a convincing video that shows you saying it and there aren't any good tells that it's fake how do we govern in a post truth world? How do we trust in a post truth world? I've been bracing myself as an insider in the industry for these days since about 2017. And I know there's some great experts that are working in this space around digital watermarking and ways to identify when things have been AI generated, but I'd like to see that work invested in an amplified. I think Dr. Hani Farid is one of the world's leading experts out of UC Berkeley in this space, and he just open sourced some coursework on YouTube about how to identify that. But I think the answer to your question. Why do I go back to video? And it's because I think that's what puts us most at risk for not being able to prove authenticity.
Carla King:A lot of writers do use video, and in fact, before we started rolling, I was talking about a podcast or a video AI tool called Descript, D E S C R I P T, which I just love. and me being ethical. I could take your voice and make you say some interesting things, right? And cut and paste from our session here and maybe take things out of order, et cetera. But I use it for taking out The long silences when we're thinking or the bloopers when the dog barks or the too many uhs and that kind of thing. And that's a great, that's a great tool. And that's a tool that I think a lot of authors and business people can use.
Lisa Thee:And what a great example of an ethical use, right? Your area of expertise is writing. You enjoy interviewing us on podcasts, but video editing is not necessarily the area that you spent the majority of your time investing in as a tool, right? So you're allowed to use more of your time for your subject matter expertise, which is your writing and speaking and less time doing the technical editing of videos because your AI assistant is working alongside you to help you recover that time, but still get a professional outcome. And I think that's a great application example of what a lot of us will be experiencing in the writing world or in the offices as well.
Carla King:I have to say that it really has freed me to edit the podcast episodes myself because sometimes I have to cut 10 minutes, and I know that, but if I had to tell an assistant, a video editor, how to do that, it would take a lot of time. This tool lets me actually highlight In the transcript what I want to take out and it syncs it with the video. It's just an amazing way to be productive and to use AI for a marketing purpose to start a podcast or a video series or a webinar series or an online course or whatever and make it look very professional and make some money with it.
Lisa Thee:Absolutely. And it's a low cost solution, right? Having to go on Upwork and find a video editor and pay them by the hour is a barrier for entry for a lot of people. And so being able to leverage some of these tools to recover your time and your talent, I think is a great application. And the video series that I host we use some of the more recent video editing tools available through Adobe and other providers, and we were able to go from from five days. Post filming for a full edit of our video series down to one day using the new AI tools. So imagine having a whole video production team saving 75 percent of their time in post production and being able to go film more things and do more interviews.
Carla King:And you probably need to pay fewer people to do that for you as well.
Lisa Thee:Yeah, which is like the pros and cons, right? It's awesome if you're the writer that needs the marketing material, not so awesome if you're the video editor that has been doing that professionally. So I think it's going to be an opportunity for us to All become a beginner's mindset again and see what are our unique values that we bring with our jobs and looking for ways to amplify that and that's actually one of the inspirations of why I wrote Go! Is I think in 20 by 2025, we're all going to have to upskill and be more data literate. It's going to be a baseline requirement. For being able to be successful out in the world as AI continues to advance. And so I wanted to give people a framework for kind of sussing out what got you here won't get you there. And how do we identify what are your unique strengths and talents that you bring to the table. What inspires you? What's your vision? What makes you uniquely human? What makes you thrive? And how do we shift you from maybe spending half of your day doing that to spending 80 percent of your day doing that? Because you're able to uniquely do something that Makes you irreplaceable
Carla King:And it does give you more time to practice your area of expertise, right? I like to say, I produced the podcast in half the time I can go for a motorcycle ride, and then I could go travel more. I can, work three days and travel for a long weekend,
Lisa Thee:and for a wonderful audience that is looking forward to hearing about all your adventures, they want more time for you in the seat of that motorcycle creating your next adventure versus producing a podcast to talk about it.
Carla King:And writing the book, and I've found also that AI is wonderful for outlining a book, for instance.
Lisa Thee:Yeah, I would say that the tools that are available today weren't available, at least that I was aware of in 2021 when I started outlining my book and I relied on humans to do that developmental editing and as a first time author, I would not have gotten through the process had that not existed. So having these tools that democratize that access where you don't have to invest as much time and money to see if you think your idea really has something that's ready to go to the next level. I think it's wonderful because it's going to amplify more marginalized voices that haven't had access to being able to share their message before. And if there's anything that I hope for in this post AI world, it's that we can take a step back and get a broader perspective on things and hear from more voices and integrate that because we have the ability to ingest all of that information.
Carla King:So are you saying that people who are marginalized and have not had a voice in media their work is being consumed by AI or they can use AI?
Lisa Thee:I think more on the second, they're going to have more access to resources to getting their message out there. For example, if you live in a place that is not a native English speaking place and you are a prolific writer, but it's been locked behind a wall because of the inability to translate it into a language that is more spoken around the world. Perhaps, using an AI translator tool that have been growing in terms of effectiveness before they were pretty clunky because they were trained on a lot of like medical sounding data, very formal speech kinds of things, and it just didn't feel fluid, but I know that companies like Meta and Google have invested a lot into translation services and AI to make it more authentic and more genuine. And so I think of the writer that's perhaps in a far away place that has these amazing tales to tell that wouldn't be accessible for the world had it, had we not had some of these technology capabilities.
Carla King:And there are a lot of AI editing tools now as well. I can just name three that I love. I love ProWritingAid, and there's Autocrit, Fictionary, Grammarly.
Lisa Thee:Even my children are using Grammarly at this point, and I think how wonderful. I did not get comfortable with grammar in elementary school. There was a couple gaps in my learning maybe transferring schools. I missed a section or so. And so I didn't even attempt to write until I was 45 years old, because in my mind, I was bad at writing and so I think that with these kinds of things that can help teach by example and help us to get past some of our maybe barriers that are driven by shame or embarrassment, it can unlock more voices.
Carla King:Yeah, I get it. I was like that with math, but later discovered that I'm really good at math. Go figure. AI can help with all that, we'll just go back to working with AI. What I've coached writers to do, especially authors who are writing articles to promote their books, is to use prompts to create lists of possible articles and then also use prompts to outline the article and then again to suggest rewording or reformatting for a beginner's mind or for a professional, like this article is for Psychology Today magazine for people who are in the psychology profession, or this article is for somebody who's just starting to study psychology.
Lisa Thee:I love that because as somebody who's worked in a field for quite a while, you get so immersed in the jargon after a while that you don't even realize that you're speaking jargon and that people can't connect with it.
Carla King:Is there any last advice that you have for us or tools or resources that you like to share?
Lisa Thee:Yeah, sure. So I think it's important to recognize that AI is only as good as the data that it's trained on. And you do have to keep out an eye out for a few things in your final final drafts of whatever you put out there. Be aware that AI is prone to hallucinations. What that means is it just makes stuff up and it sounds like a very shouldn't competent, confident delivery, but that it might be complete BS. So you want to make sure that you're using your expertise to fact check whatever it produces that you are going to put your name on for marketing purposes or whatnot. Second is that you are giving away your information. If you're using a free tool, I know I said that earlier, but I'm going to double down on that, which is do not put anything that is Personally identifiable information in there. Do not put anything in there that you would not want splashed over the page of the New York Times. You have given permission for your information to be. Used for many different applications and trust me just in the as in the world of social media. If you are not paying for the product, you might be the product. So be aware that there's a trade off happening there. And last but not least bias in tools can be really problematic. The reason for that is that AI is trained by humans and it also producing outcomes based on how things have worked historically. So any kind of embedded systemic prejudices or challenges that we face as humans will be replicated in terms of the recommendations of what it produces going forward if it doesn't have data to counterbalance that so be aware that this is a great rough draft generator. This is not a great final draft generator and when models are built, they're usually built for a very specific purpose. We are not in the age of generalized models today where you can just apply them blindly to different places and not have some unintended consequences like we hear the new cycle of people being negatively affected. An example of that is Google used AI to predict who they should hire out of a hiring pool based on who they've seen promoted through the ranks over the years looking at their own data set. In tech, it's been a known problem that there is some gender discrepancies between who gets selected for promotion. It literally did not recommend to invite a single woman to an interview on the first pass. So that's an example how bias can cut people out before they even have a chance to compete. And of course, they went back and started tuning those models to give more information about women that have thrived, and then the model was able to give better predictive analytics for what we want to see going forward. But just be aware, it's going, it's trained by humans, it's the data is curated by humans and the patterns of what it's predicting, all it's going to say is this worked before, so we think it'll work again. If you're trying to change society towards a more forgiving and kind society where more people can thrive, you're going to have to be really intentional about making sure the models are seeing examples of that.
Carla King:Thank you for all the really wonderful work and advocacy and watchdogging that you're doing with ethical artificial intelligence and society and it's really wonderful.
Lisa Thee:Thank you, Carla. I never would have believed in a million years when it was coming back to tech after having my children that I would be in one of the areas that's in the news cycle, so to speak around technology, but I fell more in love with a problem and apply technology to address that problem, which is probably a little bit different than maybe some other people have come into AI as it's becoming really. Widely known for me, I would like to see a world where marginalized women and children have better protections by bringing. Influential leaders and innovation together to address their needs. And so that's how I choose to apply. AI for AI's sake isn't that interesting to me, to be honest, but I would love to see more voices unlocked in the world to see different perspectives from the writing community because they have more access to the resources for getting the editing done, getting the publishing done, getting the marketing done, that can be really cost prohibitive and so I think there are a lot of ways that this can really bring populations up to more opportunity. And that's what I hope we can see.
Carla King:Thanks so much, Lisa. And to our nonfiction author listeners and the professionals who help you succeed. Remember, keep writing and publishing. The world needs your experience and expertise. Thanks for listening to this episode of the Nonfiction Authors Podcast. You can find the transcript, show notes, and links for this episode on the courses and events tab@nonfictionauthorsassociation.com. While you're there, please check out our membership options. We've negotiated discounts for our members that can more than pay for your membership fees, such as 20% off editing and 10% off audiobook production, 33% off our master courses, discounts for printing, publishing, and website development, and discounts from Office Depot as high as 80%. We send out an author advisor email every Friday. That includes media leads to everything from periodicals to podcasts where you can promote your book. Members can also participate in our very popular mastermind group. We call the Author Brainstorm Exchange, and you'll get access to our private community too. We have a year round nonfiction book awards program, an annual nonfiction writers conference. Classes and master courses on book marketing, book publishing, and book promotion, all of which come with optional professional certification for author assistance, consultants, coaches, and others who help authors succeed. There's a lot more to explore@nonfictionauthorsassociation.com. We hope to see you there.