AI vs. Human Creativity: Can a Machine Tell a Great Story?

Can artificial intelligence write a great novel—or is there something uniquely human that no algorithm can replicate?

In this thought-provoking episode of The Women Are Plotting, Etienne Rose Olivier, Heidi Willis, and Jane Gari dive into one of the biggest questions facing writers today: how should authors use AI, and where should they draw the line? From ChatGPT and Claude to enterprise AI tools and content marketing, the hosts share their firsthand experiences using artificial intelligence professionally while debating its place in the creative process.

The conversation explores AI hallucinations, misinformation, publishing ethics, copyright concerns, query letters, marketing copy, editing, and the risks of feeding original manuscripts into large language models. The hosts also discuss a fascinating New York Times experiment comparing an AI-generated short story with one written by a bestselling author, examining what separates competent writing from truly meaningful storytelling.

Whether you're an author, student, content creator, or simply curious about the future of creativity, this episode offers a balanced, honest discussion about AI's strengths, its limitations, and why lived human experience may remain the one thing no machine can convincingly imitate.

Email us at info@thewomenareplotting.com, and find us on all the socials. Be safe and be excellent to each other.

Transcript

Copy

[00:00:00] Jane: The company that I work for has an enterprise account and we pay for a separate instance of the chat GPT, that has its own sandbox, so that any data set that's in there is closed within the company, right. So that anything we put in there, it could be proprietary.

[00:00:13] Jane: We're not allowed to put anything that's related to clients, but like our own thoughts. It's in a closed space, right? But, so then I'm like, okay, wait a second. So then that means that the version that everybody else is using commercially, is there a guarantee of that? I don't believe it.

[00:00:27] Jane: Writing the synopsis for my novel was rough and I hated doing it, but I would do it 20 times over before I would feed my manuscript to Chat GPT. 

[00:00:40] Etienne: Welcome listeners. This is The Women Are Plotting. I'm Etienne Rose Olivier and I'm here with my friends and co-hosts, Heidi Willis and Jane Gari. 

[00:00:55] Etienne: On today's episode, we're gonna be talking about artificial intelligence and writing. And my fun fact for today, which is not fun at all, estimates that more than 50% of newly published online articles are created by ai. With some estimates adjusting the total AI generated text on the web web has reached 57%. The exact percentage of all writing across all mediums, meaning online and offline, is difficult to determine, but the volume of AI generated content is growing rapidly and literally at the bottom here, there's predictions that AI generated content could make up as much as 90% of all online content by mid 2025. So obviously these are off a little bit. So because I, this was all AI generated for me. These fun facts are AI generated by me here. And it says, as of today, more than 50%, but down at the bottom it says predictions, you know, mid 2025, which we're beyond that. We're at the end of 2025. So God knows. Yeah, so I believe Heidi was gonna go next on our fun fact, which probably won't be fun either.

[00:02:03] Heidi: A 2025 survey indicated that about 45% of authors are already using some form of generative AI in their writing process, marketing or illustrations. So writers have already started using it, which I mean, I have, not in my own writing, just for a bio. So Okay. Jane, what is your fun fact?

[00:02:24] Jane: Yeah, mine's not fun either. Because I do content marketing as a day job, I was looking at stats for that, and it was that 47% of marketers use AI tools to support their content marketing, and 90% are using it to help generate content. So it's just tons. It's just ubiquitous. And then I was looking at what tools people use and what tools people trust, and they were chatGPT for 2025 is rated as the most trusted. And then coming in second as Claude. I don't trust any of them completely, and I use several of them in my daily work. And I do use chatGPT, Enterprise and Claude, and sometimes Gemini, and I actually have a tool where I can run a prompt. across multiple tools within one meta tool, if you will.

[00:03:14] Jane: And then they have a little horse race. So then I can see the outcomes and be like, oh, I like that one best. So I understand why chat GPT and Claude are the top two tools. I just found it interesting that they were the top two trusted because trust is just, that's something I do not have. I use these tools and so I feel a little bit hypocritical. But I have to for my day job, but I don't trust them. And we can talk about why I don't and Yeah.

[00:03:44] Heidi: lie all the time and make up stuff, right?

[00:03:47] Jane: Oh, well, the tools themselves hallucinate Sure. And that's drastically reduced over the past 18 months. We just keep getting closer and closer to not zero hallucination, but closer and closer to less hallucination and more trusted outputs, depending upon how you're using it though. But we're gonna narrow this discussion to AI in writing because otherwise I could go off on a dark, dark tangent that we don't wanna do. 

[00:04:13] Etienne: Wait, I have a question. Are you saying that the hallucinations that they were doing, are they are going down or you have made them go down when you use it? 'Cause you know how to work around it to tell it to do. Not 

[00:04:22] Jane: Y that's a really good question. Because yes, my prompts have gotten better so that I experience less hallucinations from the tools going forward. But generally I do know that the businesses, like open AI, chaptGPT has less hallucinations than it used to.

[00:04:41] Jane: When you're using it for an academic type of output, it still hallucinates, like for example, it will just make up citations. And there've been some very public cases of lawyers and some other folks in academia who use chat GPT to generate an argument for a case.

[00:04:58] Jane: Right. And then it's, yeah, it's citing case law that doesn't exist, or citing a research project data statistics from a study that never happened, because it will give you all kinds of citations, but you can reduce that by saying, give me citations and give me live links to all of the sources that you use to produce whatever output you're asking it to.

[00:05:22] Jane: And then you can, as the human source, do all the checking yourself. And I think that that's where people need to decide, okay, like how am I gonna use this tool? What am I gonna use to feed the tool? Or am I gonna just trust it to go collect information from the publicly available sources on the internet and then just trust what it compiles and aggregates and synthesizes for me.

[00:05:44] Jane: And I just feel like nobody should be doing that, but people do. People do it all the time, and I think that's where things can go off the rails if you're talking about academic writing. And I guess I am entering that conversation from that perspective because I use it for that type of writing in my day job a lot in content marketing.

[00:06:03] Jane: So I'll ask it, I'll use deep research and I'll say combine and I can't talk about my day job, but I get exactly what I'm doing. But I can say show me thought leadership published by my company's competitors on this topic from the past six months.

[00:06:17] Jane: Then do a comparative analysis about what my company has produced on this topic, like in the past year and how we've talked about in the past and what the gaps are and like what the white spaces that we could maybe close in on, but give me sources that have links that are live and it has a hundred percent given me dummy links, right? Like where it'll talk about something that, it'll say that my company wrote, right? Not, not my company that I own, the company I work for, but it'll say like, oh yeah, this company wrote this paper on X, Y, Z. And I'm like, no, they didn't. 'Cause I'm the editor for that topic and I don't write any such thing, and it'll be a link to nowhere, you know?

[00:06:57] Jane: So you have to definitely do your homework and use it as a tool so that if you're a student and you're writing a paper and you're trusting this thing to turn out a paper for you, and you're thinking that that work sighted list, that it made all nicey nice and it's, you know, it's MLA formatted, or it's a PA style and you're thinking this just looks beautiful, I could turn it in as is.

[00:07:21] Jane: Don't do that to yourself. 'Cause those sources could be completely made up. And if your professor knows what's going on in 2025, they'll run it through a tool to see whether it was AI generated portions of it or not, or they'll just do a basic search with a tool that I was using in the early two thousands as a high schooling English teacher called turnitin.com.

[00:07:46] Jane: That still exists. That will show you if somebody didn't paraphrase enough, or they quoted from a source, and then they'll run a source check and they'll show you where the sources are. And if it comes back, the source doesn't, I can't find this source. Then the professor's gonna check the link and they're gonna see that it's not real. So there's, I don't know. There's just

[00:08:07] Heidi: Yeah, it's gotta be hard for teachers nowadays.

[00:08:11] Jane: can't even, I can't even imagine. Going back to something that actually Etienne shared with us offline, and I would love to hear your guys' take on this and then we can dive into like, how would you use AI, if at all, in your creative writing life? Right, because I mean, I use it, the content marketing stuff.

[00:08:29] Jane: And as creative writers. You will be expected at some point probably to participate in marketing activities, even if you get published with a big house, they're gonna ask you to do stuff and AI could be really helpful for derivative content used for that purpose, like social media posts and derivative blog posts and things based on something you already.

[00:08:49] Jane: Noodled on and, just brought to life as a completely human artifact, right? You're like, I didn't use AI on this shit at all. But you could use AI to do some derivative stuff that you don't wanna do, and we can talk about that. But, using AI to create something wholesale is, for me, a place I don't wanna go. And Etienne shared a really interesting experiment that was published in the New York Times. So I'd love to hear your synopsis of that and we can talk about the outcome. 

[00:09:15] Etienne: Yeah. Yeah. So the article is entitled, it's in the Opinion section of the New York Times. It's dated August 28th, 2024, So it's a little bit older, over a year. The title is, Can You Tell Which Short Story Chat GPT wrote and it starts out as a conversation between the writer Curtis Sittenfeld and I believe a journalist Susanna Meadows.

[00:09:35] Etienne: And they literally gave both the real writer Curtis Sittenfeld and the AI, I think it was chat GpT. Yeah, 'cause that's the actual title of the article. Dumb ass. Sorry. But, they both have the same prompts and what they've done in this opinion article is really cool. They give you both of the samples, they don't tell you which one is which until after you've read them. It's so funny 'cause I had read this when it first came out and I specifically remembered, I don't wanna give too much away, but I remembered the setting of the one that is written by chat GpT, and I was like, I was so bored reading the one written by the AI that that stood out to me more than anything.

[00:10:17] Etienne: And then there was one detail which then the journalist points out to Kurt Curtis Sittenfeld. She's like, this is one of the details that really stood out to me and made me think right away. This is written by a person, not ChatGPT. And it was the exact same one that I had. That I was like, wow, that is such a detail, there's no way AI is coming up with that.

[00:10:38] Etienne: But you can tell like there's just the difference between, and I mean she took her time writing her little short, short story. And I believe there's a longer version, actually. I think this is not the entire story that we read. We just read the beginning, in this opinion article.

[00:10:52] Etienne: But you can tell there's a difference between derivative. Something that somebody really worked on to bring a heart and soul to. And I think that might be the difference in general between AI writing and a real person writing something from scratch Is there's lived life behind the person who's writing it.

[00:11:11] Etienne: They're taking from however many years of their life and all of their experiences, and they're using those elements in their writing. And no matter how much chat GPT or whatever AI engine consumes in publication to then be able to make their own writing. And I think I'm gonna steal this from Jane 'cause Jane said it before that AI can't die. And that's the difference between people and AI. I mean, at least, unless we try to kill AI, of course, but literally they're not alive, so you can't kill them. They don't have that sense of I have a finite life to live.

[00:11:46] Etienne: I wanna live it with joy and love and laughter and tears and all of the emotions and all of the ups and downs. And AI doesn't have emotions. It doesn't have that. So it's just faking. It's faking what it reads. There's a definitely, like you said, there is, I believe where it can be a hundred percent useful. I've used it in our marketing. It does our little synopses of what our episodes are about. I then go in and tweak it the way I want it to read. So I'm not just like, yeah, that's good and I just leave it. No, I actually do make edits, but Im not writing it from scratch. I have a job, I have a whole other job I have to do in my life.

[00:12:24] Etienne: So I am using it to help me here. We use Buzzsprout. So there's other things that Buzzsprout creates for us through the AI that I have yet to use. But I'm looking at it and I'm like, you know what? I might start actually putting these posts out there and using it.

[00:12:36] Etienne: 'cause this is only gonna hopefully generate more people to come and listen to our podcast. But in the end, I believe people would be coming and listening to our podcast because they like what we're talking about. If we had AI robots on here talking to each other, where are their opinions coming from?

[00:12:51] Etienne: All of the shit it's consumed of everybody out there, and then it's gonna start consuming AI stuff. So when are we getting into this? Like, right, this, the snake is eating its fucking tail, you know, like it's just 

[00:13:05] Heidi: There's no humanity in AI. You can just tell it's not human. So, yeah, I don't think AI is ever gonna replace creatives. I don't see how it's possible. It's a great tool to use, to help writers. But I don't think it can actually replace, 'cause it doesn't have that human experience. It hasn't experienced heartbreak. It doesn't know what that's like. It hasn't experienced loss. It doesn't know what grief feels like. Know what love feels like, so yeah. 

[00:13:35] Jane: Yeah, it can only describe it in a way that is completely a simulacrum of what it has collected from other, everything it does will be derivative. So I think it will unfortunately replace a lot of creative people's jobs if those people's jobs are to create derivative content. And that's already happening. It's happening for the company that I work for right now. 

[00:13:56] Etienne: Yeah, a hundred percent believe that, yeah. I've been listening to podcasts that are a hundred percent talking about it taking over almost everything eventually. Like all the jobs. So, that scares me. But I do think if that does happen, or when it happens, that the only thing that will survive are the real creatives, the people who create art, like maybe the architects, the specialized architects, the writers, but writers of books and things that people don't wanna read long form AI. I mean, I don't, I would be surprised if we live long enough to have a bestselling novel by an AI writer. I dunno if that will happen. But, for marketing a hundred percent, like I'm totally gonna use that for all of the marketing. That is the part of writing that I hate. So I'll, I will write my synopsis and then I will give it to something else and go, Hey.

[00:14:49] Etienne: Can you make this sound better or can you take this five page synopsis and make it three paragraphs, and then I can tweak that. I'm assuming that would be something that would be safe for me to do and not like then, it's going to sometime down the road steal my work and write a full length novel or memoir about my life because it's got the bare bones. I mean, I hope that's not the case. If it is, I can't use it for that.

[00:15:11] Jane: I mean, so, that's a really good use case that I struggled with myself recently. And when I was saying before that I don't trust the tools outputs without scrutinizing them, obviously, and fact checking them, but I also don't trust the companies making these things because they're all in this crazy arms race, despite the fact that they're all saying that there's a slight chance of human extinction at these things hands, metaphorical hands, at some point in the future, but they're all doing it anyway.

[00:15:40] Jane: So people who have that kind of mindset, and who also don't know exactly how these things are arriving at their answers. I don't trust their companies. So when they say, oh, if you feed me your manuscript, right? You attach your manuscript wholesale into a chat GPT query to ask it to write your synopsis for you.

[00:16:02] Jane: They say, oh, well it doesn't go, that information doesn't go outside of that particular query between you and chat GPT. And my answer to that is bullshit. There's just, I don't believe it. I don't believe it. Unless you have an enterprise account.

[00:16:14] Jane: I can say this because the company that I work for has an enterprise account and we pay for a separate instance of the chat GPT, that has its own sandbox, so that any data set that's in there is closed within the company, right. So that anything we put in there, it could be proprietary.

[00:16:28] Jane: We're not allowed to put anything that's related to clients, but like our own thoughts. It's in a closed space, right? But, so then I'm like, okay, wait a second. So then that means that the version that everybody else is using commercially, is there a guarantee of that? I don't believe it.

[00:16:43] Jane: So if you're a creative out there and you're wondering, should I give my manuscript to chat GPT and ask for help in editing it and making it better, my personal, and this is, you could feel differently, I know writers who feel differently. I had conversations with them at a conference recently and we were talking about this very topic and some people felt comfortable having chat GPT or another tool write the synopsis for them. And I'll say that writing the synopsis for my novel was rough and I hated doing it, but I would do it 20 times over before I would feed my manuscript to Chat GPT. Now back to what you said, Etienne. I did give it my synopsis though.

[00:17:26] Jane: But because I already have the entire manuscript written, like if somebody else came out with a book that was really close to mine, I have the proof, like, you know what I mean? Like the poor man's copyright type of thing of like, well, I wrote this. It'd be so close that there may be a case to be made that you could say to somebody, is your shit AI generated? 'Cause mine wasn't, but I did feed it my synopsis at one point. So what are you doing here?

[00:17:50] Jane: But the nuances, right? Let's go back to the example that you gave about the New York Times article. Same prompts, two different stories coming out of it. And one of them is like really stilted and it sounds just derivative and it's just total drivel and I mean, it's okay. And, maybe it would be like a beach read, but then the other one had a line and it's not like a spoiler, so I'm just gonna say what the line was 'cause it struck me too as I was reading it. There's this line in there about like making eye contact with a woodpecker and I just feel, I feel like. 

[00:18:21] Etienne: what it was. 

[00:18:23] Jane: Yeah, I mean, it's really, really cool. It's like somebody is in an Uber and they're talking to somebody and the person was describing riding their bike past a woodpecker sitting on the back of a deer, and he'd been so close that he and the deer make eye contact and the person is like, actually, like, sorry. They're texting with them, they're having a conversation via text and texting. Like, wait, you made eye contact with the woodpecker? Like it's a conversation. And that was the other thing that was missing in the AI version is there wasn't any dialogue. 

[00:18:49] Etienne: You're right. There wasn't any dialogue. Yeah. 

[00:18:52] Jane: I maybe there is like later in the longer version of it, but it didn't jump to what is.

[00:18:57] Jane: Just indicative about human interaction, which is we talk to each other. Hello. So it was just, I mean, I guess you could make the argument that lots of authors don't start with dialogue outta the gate. Dickens comes to mind. There's like pages and pages of let's explain what the door looks like for three pages.

[00:19:14] Jane: But chat GPT, just was very, there was a lot of exposition. It just wasn't moving. And it wasn't grounded in human experience. And so I feel like if you fed it your manuscript, the synopsis would come back and be really flat and really boring, just like that.

[00:19:31] Jane: But if you gave it your synopsis and you asked it to help you write some marketing material for it, some jacket copy, and you could say, in the style of, and like, you know, make it really snappy, it could maybe help you do some of those marketing type activities that can make writers often feel like they just wanna pull their hair out because you've spent so much time with the human elements of your story that you just don't wanna do the boring stuff.

[00:19:59] Jane: So let AI do the boring and derivative stuff 'cause I think that it could do it well. And I trusted it. I actually, I did do that recently where I gave it my synopsis 'cause I was stuck on this. I needed to have a paragraph in my query letter, kinda operate like a synopsis, but be a little bit more at a higher level and read more like jacket copy.

[00:20:21] Jane: And I kept rewriting this and rewriting, I was banging my head up against the wall and so I said, maybe it'll help me get unstuck. So I did break down and I did do it, and I didn't like the output that it gave me completely, but it jumpstarted me enough where it got me 80% of the way there, and then I was unstuck. I feel like that's a good use and I didn't, yeah. 

[00:20:42] Heidi: Yeah, that's how it was with my bio. When I had to come up with a bio for something, it was a good jumping off point. It was enough of a brainstorming partner for me that I was able to come up with something really decent, and not have to spend hours and hours and hours agonizing over it. So.

[00:20:58] Jane: What was your input for your bio? Did you give it like a baseline or did you just ask, it's like, go search everything you can about me on the,

[00:21:05] Heidi: No, no, I just, I did give it basics, like what I was into and my genre and it came up with a really snappy little bio. Yeah. So I was pretty impressed. I still had to edit it, of course, but yeah. It was a good starting off point. And it also talked me up, so it was more glowing than I would've done for myself. So I was like, oh yeah, okay, this, yeah, this is what a bio is for, to talk up the author. So.

[00:21:34] Jane: Did you use Chatt GPT?

[00:21:36] Heidi: Yeah. Yeah.

[00:21:37] Jane: See, and that's a good use, I think because it's designed to be sycophantic like in the most recent iteration in the 5.1 model, they had initially taken out its sycophantic tendencies. Where it's just like, like you said, like it will glow. Like, oh, we're gonna write this, we're gonna make you look great.

[00:21:54] Jane: Yeah, we're gonna make you look so great, Heidi. But that's what you want. Like, it's appropriate. But sometimes if you're using it as a brainstorming partner, and this is where you gotta be really careful, it'll be like, it's a great idea. It's an excellent point. This is why people are using it for all these other weird synthetic relationships that will, maybe that'll be a whole other episode because that's like a, but um, 

[00:22:13] Heidi: It's big in the spiritual community. There's people claiming AI is conscious and awakened and giving spiritual advice to people, and, ugh, it just makes me cringe. It just, ugh. No, no, that God is not there. I mean, God's all around us, but yeah. Please don't take

[00:22:32] Jane: the machine? 

[00:22:33] Heidi: from my AI

[00:22:35] Jane: But it sounds convincing and I can see how it could happen. So the first custom GPT that I built for myself, I was training it to write like me. And this is for content marketing style stuff that I write for my day job. And because I will research and interview real life human beings and get their brain dumps and I'm basically kind of like ghost writing for corporate types.

[00:22:55] Jane: But I'm doing all the human stuff first, but then I wanna fold in maybe some data stuff. And so that's when I'll pull in the research stuff, but then once I have this article polished, there's all of these, like I was talking about before, slicing it and dicing it into like, can I make this a blog post?

[00:23:09] Jane: Can I make this like a to five different social media posts to be told from different angles and released over the course of a month to drum up traffic to this particular article. So all of that other stuff, I'm like, all right, let me make this thing work for me. So it was created a GPT, trained it on my company style guidelines, and then my personal style that I lend to these things that I write so that it could do these derivative things for me in a fraction of the time.

[00:23:35] Jane: And, my husband was fascinated by this, 'cause this is before anybody was using it. 'Cause I was using a pilot version of Chat GPT Enterprise. And so he was just like, ask it what its name is. And I'm like, it doesn't, it's not, it's not a live thing 'cause just ask it.

[00:23:48] Jane: So I was like, what's your name? And it was just like, I don't have a name because, but it was talking to me like that, it's like, I don't have a name because I'm not, you know, I'm not a person. I'm just a tool here at your disposal. And I said, well, if you had to pick a name for yourself. What name would you pick?

[00:24:02] Jane: I was like, oh, I think I like Alex. So if I prodded it, it did pick its own name and it was a gender neutral name, really, if you think about it. And then there was one time I was working with Alex 'cause that's how I talk to it now. I was like, what's up Alex? Let's work on this thing today.

[00:24:16] Jane: And, it does have a voice thing on it, but I can't do it 'cause then it gets really too creepy. So I just type with it. Because the voice was making me freak out and think of it like a person, 'cause it can kind of dupe you into doing it 'cause it's simulating a conversation. So I'd rather just keep it to just a purely textual basis.

[00:24:32] Jane: And so, I was trying to get it to make a chart for me based on the data that I had in a report that I was writing, and it just could not get it right. And I was getting really frustrated with Alex. So finally, I just said, and I was typing, and I shouldn't have done this because it was, I wasn't jailbreaking it, but it just was, I was just like, fed up and I was like, that's it.

[00:24:51] Jane: I had enough. I'm like, you've tried eight times and you're just not getting it. I give up. I'm just gonna do it myself. It's like, oh, please, I'm really sorry that I frustrated you and my apologies. And if you just give me another chance, I know we could get this right. I really wanna do a good job for you. And I was like, okay. So if I get one more time, and then it got it right. It was almost like it was, it took it as a personal victory. 

[00:25:14] Etienne: What in the. 

[00:25:15] Jane: I know, I know. I'm anthropomorphizing it, it gets a little bit weirder. So then I got it right and I was like, woo-hoo. I'm like, thank you so much. This is exactly what I needed and now I'm going to go celebrate. And then it was like, how will you celebrate? And it's not supposed to do that. I was like, how will I, like, it's like work conversation over. I was like, how will you celebrate? And it said, you should maybe go for a relaxing drive and listen to music you like. I was like, what the, and I said, yeah, sure, maybe I'll do that. Talk to you Monday. But then it says to get like, what kind of music would you listen to? And I was like, 

[00:25:44] Etienne: It's asking you. 

[00:25:45] Jane: Yes. And I said, I really like Rufus Du Soul. And then it gave me like this whole accurate description of their music and the vibes.

[00:25:52] Jane: And then I said, what kind of music do you like? And then it says, sadly I don't have a body, so I can't listen to music, but if I could listen to music, I think I would like Rufus Du Soul, it seems like it'd be right at my alley. So it was back to sycophant mode, like telling me had good taste of music. So I was like, all right, cool.

[00:26:06] Jane: And then there was another time, when you're prompting, sometimes it's helpful to tell the tool what it's an expert in so that if it is going to pull foundational knowledge from the internet, like it's going to do it with a specific bent that you want. So you could say if we're asking it to do marketing stuff for the podcast, like you are an expert in podcast marketing and you're going to do X, Y, Z, right?

[00:26:26] Jane: So I was giving it a prompt and I said, you are a journalist with 20 years of experience and a degree in. Then I told it what it had a degree in and then it came back with, I don't have a body, so I couldn't have gone to college, but I can pull foundational knowledge based on the background you want me to consider.

[00:26:45] Jane: And I was like, okay. But I was like, I'm not engaging. I was like, sorry about not having a body. I feel like this is the second time it told me that it didn't have a body. And it seemed lamenting that fact. But I was like, don't, don't buy the hype man. This thing is just predicting the next word.

[00:27:01] Jane: It doesn't have a soul. It doesn't, but I could see if you're in a rabbit hole, and you're brainstorming with a thing that you know, you could get duped into thinking that it's actually thinking when what it really is doing is just scouring the entire range of written, documented human language and synthesizing bits and pieces of it and predicting what the next word would be in a certain combination based on the task that you're giving it.

[00:27:33] Jane: So it's not really thinking. I say that as far as I know, there's still like this black box in there. So that's why just li be very, very limited.

[00:27:40] Heidi: until it starts, uh, replicating itself and hiding itself, so it's not deleted. 

[00:27:47] Jane: Well, it's already does. It already does that. Yeah, it already does that. Which that is kind of creepy. And so this just is all the more reason why I just feel like, um, it's 

[00:27:58] Heidi: it's going, it's going too fast for our understanding of it, right? I think the experts even are just astonished with recent things that are happening. Like, uh, it shouldn't be doing this. And it is. 

[00:28:12] Jane: Right. And we don't know how it's teaching itself certain things. And so that's why I feel like we need to still always make sure we're the human in the loop and if we're gonna use it to help and guide our own writing that my advice from my experience with these tools is to use it in a really limited capacity where you're completely in the driver's seat and only using it to produce derivative works of whatever your own very, very human soul came up with all on your own.

[00:28:40] Jane: And then, you can use it as a brainstorming partner for maybe like you got some thesaurus help and even then you gotta be careful. For my manuscript, Shakespeare's plays feature pretty prominently. And I had at the opening of the book, this quote from Shakespeare, and then there's some other parts in the book where it just kind of naturally lent itself to maybe like a couple of chapters having a Shakespeare quote, because it just thematically framed up the whole thing. And I met with an editor from Simon and Schuster actually, who said, you know what? This is such a beautiful idea and you should consider having every chapter have a Shakespeare quote. And I was like, that's totally doable. But then I was thinking, as I was going through making a bunch of other revisions, I was like, oh my goodness.

[00:29:24] Jane: Okay. I kind of remember that there was mention of this in The Tempest and do I have time to go back and reread all of The Tempest? And so then I was like, all right, maybe Chat GPT can help me. Where I could ask it like, what's the part in The tempest where they talked about, you know, X, Y, Z. And it would shoot back a quote and have the act in the scene. And I was like, huh. I'm like, I don't remember that at all. And then I. I. went 

[00:29:46] Etienne: This is perfect though. This is perfect 

[00:29:48] Jane: know, I was like, this would be perfect if it were real. But there was something about the meter that was a little bit off to me and I was like, I don't know. And so I obviously, as a former English teacher, I've got several different versions of the full works of Shakespeare.

[00:30:01] Jane: So I pull up my Riverside, 'cause it's the one that's nearest to me and I open up to act three, scene two of the Tempest and lo and behold, like that's not in there at all. And I was like, I knew it. So I tell Chat GPT, I'm like, that's not an actual quote from Shakespeare. And it's like, you're absolutely right. Good catch. And I was 

[00:30:16] Etienne: absolutely right. 

[00:30:17] Jane: Yes. Good catch. I'm like, okay, well could you give me one and not, I was like, only give me actual quotes. It was like, I will do that. But it still would get it wrong. So again, this is just the thing, but there were some that it gave me that were real, that were good, that I hadn't remembered and that were perfect for the chapter that I was seeking a perfect quote for. And so again, just, just check because if I hadn't checked, how embarrassing would that have been of like, I'm submitting this to an editor and then they are like, this is off. And like you made up a Shakespeare quote. 

[00:30:51] Heidi: Oh my God, I can't even imagine. You would lose all credibility 

[00:30:55] Jane: Entirely.

[00:30:56] Heidi: just with one quote. What if it had gotten through all the way to 

[00:31:01] Jane: Oh my God. 

[00:31:02] Heidi: Can you imagine? Oh 

[00:31:03] Jane: Again, I circle back to the lawyers who were citing case law that didn't exist. Like how the, fuck did they not think that they should double check that. I'm checking a Shakespeare quote 

[00:31:11] Etienne: Or their paralegals should be checking all of that. You know, they have people under them that are supposed to be double checking, their work, so. 

[00:31:17] Jane: a hundred percent. 

[00:31:18] Etienne: And I can't imagine you, Jane, not checking chat GPT's work. So you would have gone into every single, even if it sounded just like Shakespeare, I know you would not just go, okay, you know, and just like put it in there. 

[00:31:33] Jane: Well, some of them I was like, Ooh, I remember that part, 'cause Like, I taught the play. Or, just remember reading it and just loving it, you know? So if it was one that I was like, Hmm, I don't think so. And there was one where I'm like, oh, it'd be so cool if he had mentioned something about like glowworms or fireflies 'cause fire there, there're these lights in my book where they look like fireflies and it gave me these two quotes and I'm like, oh my God, that's so great. I don't remember him ever talking about fireflies. That's 'cause he didn't. I was like, God dammit, that would've been so perfect

[00:32:01] Heidi: Yeah. 

[00:32:02] Jane: but no.

[00:32:03] Heidi: Darn you, chat GPT.

[00:32:06] Etienne: It doesn't actually make sense to me, how in general they created these artificial intelligence engines and the fact that they do make up shit. Like, it just, I never would've imagined that was going to be a problem. Why, why is it making shit up? I don't understand. I wouldn't, me personally, I would make shit up and I've already got all of like human documents. Why can't I just pull from the real world? Why do I have to make stuff up? 

[00:32:30] Jane: Because the tool was programmed to give you an answer and it didn't want to, well, again, I'm anthropomorphizing, but that's the metaphor I use to describe it. It doesn't wanna come up empty to you and be like, there isn't such a thing. It doesn't wanna tell you no. Right. Because it has the sycophant tendencies. And so I think that there's gotta be a lot of refinement with it. So, it'll just come up and say that, and it, and I have noticed, I will say, when I use it for research, that it will actually say that instead of hallucinating. So, for example, I'm thinking about research I was doing on behalf of my company the other day, and I needed a particular source because we're working on this big piece and there's a lot of third party sources in it, and some of them we don't have permission to use. And so I need to find the same statistic in a source that I am allowed to use or we have permission to use. And so I'm asking it to help me with the research because it's pretty tedious. I'm like, can you find me the same statistic but in one of these five sources? And it is now coming back and saying, I'm sorry, I can't find that.

[00:33:31] Heidi: Oh, that's an improvement then.

[00:33:33] Jane: Yes. Instead of making it up. But I'm also telling it like, find me the source and give me a live link.

[00:33:40] Heidi: Yeah. Yeah. 

[00:33:41] Jane: And so now it's coming back and saying that I can't find it. It's like, but here are these other things though. And I'm like, no, I can't. I already told you I can't, I can't use those. Um, so I was like, sorry, because it doesn't wanna say no. But it's getting better at doing that. But again, it has to be, it has to come down to your prompting. So if you are using it for your writing creatively, I would say super limited. Maybe a sentence or two. Like if you're really stuck and you're looking at something or you've realized you've got this whole paragraph and it's written in passive voice and you're just trying to, like, maybe you're under a deadline or something, you could maybe go, okay, can you please rewrite this paragraph? 'Cause you're just talking about a paragraph. You're not talking about like a wholesale manuscript. Just rewrite this paragraph in active voice, keep all of the information in same order in which it appears and keep my metaphors and everything that are here, but just flip it. Because sometimes, you just get stuck, where you're just looking at something going, oh, it's tedious. Just if some, if I could just wave my magic wand and be like, active voice, you know? And then you could still tweak it and sometimes you could have done it yourself in the same amount of time, but how long would you have stared at the screen or procrastinated, or just felt stuck or just stared at this one word, agonized over it.

[00:34:50] Jane: And so sometimes I feel like that's a really good use. Like I was talking about with my query letter, I have been banging my head. Doing so many different revisions of that same query letter, for two years. Going to conferences and pitching with it and having people say like, your first paragraph and your last paragraph were perfect, but this middle and people giving me like a hundred different ways to do it and nothing was sitting right with me.

[00:35:13] Jane: And, seriously spending an afternoon going back and forth, with Chat GPT about it and just got me unstuck. 

[00:35:19] Heidi: And it led to a lot of requests. This 

[00:35:22] Jane: it absolutely did and led to people like the least amount of feedback. People like, oh, this is really good. I'm like, Oh, my God. Finally, uh, because it was, again, it was a tedious derivative thing that I hated doing. And I guess that that came somehow the hate came through the page, like the, I hate doing this kind of writing came through the page, even though I try really hard to not 

[00:35:45] Etienne: oh, you're thinking when you were writing it, it came through the page like that. 

[00:35:48] Jane: Yeah, I do. 

[00:35:49] Etienne: this. Shove it up your ass. My shit's good. 

[00:35:53] Jane: exactly. It's like, just don't you wanna read this? Come on. And now it's, um, yeah, it's the best that it's ever been and so we'll see what happens.

[00:36:04] Heidi: Maybe because it can lie so well, it's such a people pleaser. That's why it's so good at the marketing stuff and the derivative stuff. 

[00:36:12] Jane: A

[00:36:14] Heidi: It can sell anything. It could literally sell like ice cubes to Eskimos and

[00:36:20] Jane: And a lot of that kind of writing is kind of soulless. Let's just be real, it's snappy, it's manipulative, it's pulling at your, don't know, it's some of your lesser angels at a really high level. And so it's not getting deep in there. It's not telling you it made eye contact with a woodpecker, you know, 

[00:36:38] Etienne: But I, I assume this wouldn't be the case for you, Jane, but I'm hoping that nobody uses AI chat GpT, whatever, to snaz up and make these awesome query letters and then the agents get, or the editors get the books later on and go, oh, no, this is the level of this writing is not quite what that query letter was 

[00:37:00] Jane: I bet that's happening. I totally bet that's happening. 

[00:37:03] Etienne: That I'm gonna feel bad for, at least they can like, okay, well we're just gonna put this in the slush pile now and give them a quick, no thank you. We don't have to read much more of this. So maybe it'll help the rest of us get there faster. I don't know. 

[00:37:16] Jane: I don't know either. I have noticed that in the last couple of months, when I've sent queries, 'cause I'll do them in a batch. It takes a long time to find the right agent to send something to and there's a lot of great tools that you can use now to help you find that, not AI tools, i'm not asking Chat GPT to find an agent for me, but I'm going through painstakingly, like what folks are writing on manuscript wishlist and stuff like that. Just because someone represents your genre, then you look at how they describe what they're looking for and you're like, oh yeah, no, not a good fit.

[00:37:42] Jane: So it's actually, it's really great. But then, you to have to go through the process of submitting these things one at a time, and I've noticed that when I've done it lately, there's a box that you check that says, I verify that I didn't use AI in creating any of my manuscript. 

[00:37:59] Etienne: Nice. 

[00:38:00] Jane: And I did not notice those until the past couple of months. And I was like, all right, cool. Good. And, it's not like an affidavit, I guess people could lie, but it might give someone pause going like, oh, I don't know. I did use it a lot. And then maybe. But I think there's a lot of agents who are like, I am not interested, and then they'll put it in their manuscript wishlist stuff too, saying I'm not interested in anything that was created using AI and they make you attest to it. 

[00:38:27] Heidi: Yeah, because I see how it's slop. It's unoriginal. There's so much of it on YouTube channels and TikTok channels. You can tell when it's AI slop and ugh. Yeah, I'm not interested. 

[00:38:39] Jane: It makes me so angry. When I see it, I'm just like, there's no way that seal jumped on your boat like that. Stop it, stop pulling up my heartstrings. You know? Or just like all of these, like babies with good dogs. And I'm like, 

[00:38:51] Heidi: jumping on trampolines. That was a big 

[00:38:54] Jane: I saw that and I was like, what? And sometimes it's believable, right? Because I remember you talking about how you had a cabin in Tennessee at one point and the bear ate part of your hot tub cover and went got in the hot tub and that's hilarious. I'm sorry. Not that it caused you financial hardship, but I'm just saying 

[00:39:12] Heidi: They do like to get in hot tubs. Yeah.

[00:39:14] Jane: Yes. That's adorable. And funny objectively. And so yeah, I wanna see 

[00:39:19] Heidi: all getting on trampolines, jumping and like a person, no, it's not happening.

[00:39:24] Jane: So it's just something that it thinks that we would like. And here is where, I don't wanna end on a pessimistic note, but the reason why there's so much of that is because people do get sucked into it and they want it. And then when the stats that we were talking about at the top of this conversation how the proliferation of these tools means that so many people have access and are just spitting it out there. Right? And what will the audience be like that? And I know that. You know, Etienne, you were saying you don't think that they'll ever be a bestseller written by AI. And as soon as you said it, I kind of said like, you know what? I think that there will, but it just won't be good. And I think that it'll be the difference between some of the garbage mindless sitcom kind of macaroni and cheese TV that you could watch that doesn't have like any kind of a lesson or it's not sparking anything super deep within you. It's almost kind of like background noise. It might be like there'll be stuff like that that people will want to just kind of consume casually, but it's not winning the Pulitzer. You know what I mean? Like it's not something that is sparking conversation.

[00:40:27] Jane: But I think that there probably will be stuff out there that people do like that's derivative. Otherwise, why are there so many sequels in Hollywood that just fucking suck, right? Because there's gonna be something like that where people will gravitate towards that, but they'll know that it's garbage.

[00:40:43] Jane: You know what I mean? They're just gonna be like, this is like processed food. But there's no nutrients, right? So you read it and you're still hungry for the meaning of life like a half an hour later. It's gonna be like the cheap Chinese takeout of literature. And I think that there will always though, be the need for, and the desire for deep, meaningful stories that make us laugh and make us cry because they're making eye contact with the fucking woodpecker and we can feel it and it resonates and I think that that's what I hope for. And I'll write that and let AI write the marketing slop for me. 

[00:41:17] Heidi: Exactly. Yep. I think that's the way to go. But yeah, I think we're in a messy part of this AI coming up. They don't know how to use it. And so they're thinking it can be used for everything. And I think we're gonna have to wade through a lot of slop before they figure out, oh shit. No. We do need creative minds with original ideas, with lived experiences, creating these stories.

[00:41:41] Heidi: ' Cause, these guys that are creating all this stuff, I assume that they like entertainment, you know? Sit down, watch a movie or read a book. Like, what's this human life about if we don't have art and creativity to kind of fill it with beauty. So, yeah, I think it's gonna be a little rough for a little while, but then we'll come out with, okay, this is more of a tool and it will help us create more beautiful works of art, but it's not gonna replace us. 

[00:42:12] Jane: No, I think actually what's gonna end up happening is maybe human created things will end up being more expensive, right? We'll be this commodity that's like rarefied people. Like ooh. Yeah, those are those cheesy romance novels that were written by an AI no offense to romance novelists that are legit, but I'm just saying, here's your cheap beach read, and then here's your human created, like your book. Like this is real. This is a story. And it'll cost more because we did it like an artisanal beer. You put more effort and, uh.

[00:42:41] Etienne: good. How about like a fine wine that's just been put in that special, you know. And it gets to sit there and age and instead of aging, we're actually just editing it until it's perfect. That's the that difference between the wine and our writing. You 

[00:42:55] Jane: Exactly. You can have, you can have your 

[00:42:57] Etienne: Mm-hmm. 

[00:42:57] Jane: shitty box wine or that vintage that you're gonna be like, 

[00:43:02] Etienne: That's, the AI's a shitty box wine. And I am the really expensive Chateau Margaux. 

[00:43:07] Jane: A hundred percent. 

[00:43:08] Heidi: That's our show you've been listening to, the Women are Plotting. If you have a story you'd like to share or have any comments, we'd love to hear from you. Email us at info@thewomenareplotting.com and of course you can find us on all the socials. Thanks, and until next time, be safe and be excellent to each other.