Podcast: Download (Duration: 1:10:14 — 56.5MB)
Subscribe: Google Podcasts | Spotify | Stitcher | TuneIn | RSS | More
After many years of people saying, “AI can never be creative, AI could never write fiction (i.e. make things up), it's now evident that the generative AI tools make a lot up — and we need to be aware of the potential ramifications.
How can we use the tools to achieve our creative purpose in an ethical manner, and understand that we need to curate, edit, and take responsibility for any usage? How can we educate ourselves and others on the way these AI models work? Tim Boucher and I have a challenging, wide-ranging discussion in this interview.
In the intro, I comment on ‘A concerning trend of AI-generated submissions' to short story market, Clarkesworld, and the ‘tsunami of crap' all over again [JA Konrath], and how we can use AI tools in a responsible manner.
Today's show is sponsored by my wonderful patrons who fund my brain so I have time to think about and discuss these futurist topics impacting authors. If you support the show, you also get the extra monthly patron-only Q&A audio. You can support the show at www.patreon.com/thecreativepenn
Tim Boucher is a hyperrealist AI artist and writer specializing in questionable alternative realities. He's worked professionally in content moderation policy, and counter-disinformation efforts on behalf of a major web platform, a blockchain protocol. And he has advised nonprofits and governments on related issues
You can listen above or on your favorite podcast app or read the notes and links below. Here are the highlights and the full transcript is below.
- How Tim started writing and publishing, and why he decided to experiment with AI tools for images and words
- Misinformation by humans and AI hallucinations, how we need to fact check , edit, curate, and manage outputs — and how these can be used in fiction
- Tim's AI-collaborative creative process and the tools he uses for words and images
- Labeling and ethical use of AI [see the Alliance of Independent Authors guidelines here]
- Why Tim uses Gumroad to sell direct and doesn't publish on Amazon
- How authors need to engage with the technology, experiment, and learn to stand out in an ever-increasingly crowded market
You can find Tim at TimBoucher.ca and his books at LostBooks.gumroad.com
Header image created by Joanna Penn on Midjourney.
Introduction: Addressing the flood of AI-generated content/books
In this introductory section, I want to talk about some of the news items that have arisen this week, and that fit into this episode very well.
From the headlines this week, “Sci-fi publisher Clarkesworld halts pitches amid deluge of AI-generated stories.” [The Guardian] and “Hundreds of AI-written books flood Amazon.” [The Independent].
I’m not sure why people are surprised about this, and of course, this will only increase. But as ever, the headlines are clickbait and we need to have a more nuanced approach.
Yes, there will be scammers, spammers, pirates, get-rich-quick schemes and plagiarisers using these tools to mass produce crap and publish it quickly.
But that is nothing new.
People who do this kind of thing have always done this kind of thing.
Everything I have ever written, recorded, or produced in every format has been plagiarised, pirated, stolen, and republished elsewhere. I used to try and stop it, and I still do make an effort when someone literally steals everything— but it takes up too much time to try and stop it all, so I focus on creating value for my true audience — all of you.
This happens to me, and I am a (mostly) unknown author in a tiny corner of the internet. The most famous books, blogs, films, music, etc, get pirated the moment they emerge, or even before release if leaked.
Humans are the problem, but of course, AI technology enables this to be done at scale — which makes it more of a problem.
But again, this is nothing new.
Have you checked your email spam folder lately?
Are you aware of how much content farm crap the Google algorithm filters out when you search?
Do you know how much content moderation there already is on the internet?
Are you aware of all the scams that go on even just in our little author corner of the internet?
Check Writer Beware for years of them, most of them prior to generative AI.
Do you know how many times Amazon has been ‘flooded’ with spam books?
Have a look at David Gaughran’s blog for a glimpse [Scammers break the Kindle store from 2017, Kindle Unlimited: A Cheater Magnet from 2021, and a whole load more.]
On The Guardian UK in 2018, Fake books sold on Amazon could be used for money laundering; and in 2019, Plagiarism, book-stuffing, click farms … the rotten side of self-publishing.
Because yes, we’ve heard this all before.
“The tsunami of crap” all over again
Back in 2011, JA Konrath, one of the early and most successful indie authors, wrote a blog post entitled, ‘The tsunami of crap.’
“Some people believe the ease of self-publishing means that millions of wannabe writers will flood the market with their crummy ebooks, and the good authors will get lost in the morass, and then family values will go unprotected and the economy will collapse and the world will crash into the sun and puppies and kittens by the truckload will die horrible, screaming deaths. Or something like that.”
I remember that time well.
I spent way too much time and effort trying to prove that an author could be serious about the writing craft and the business — and that professionally self-publishing, being an independent author, was a valid creative choice, and for many, a sensible business choice.
Yes, there was — and still is — a tsunami of crap, but don’t lump us all together. It is more nuanced than that.
Eventually, we stopped talking about it and just got on with writing books and reaching readers. The success of the indie author movement attracted more authors, and now, over a decade later, the ‘stigma’ of self-publishing is — almost — gone.
In fact, we are collectively a huge chunk of the book market. As Michael Tamblyn, CEO of Kobo said at Frankfurt Book Fair in October 2022,
“One in four books that we sell in English is a self-published title, which means that effectively for us self-publishing is like having a whole other Penguin Random House sitting out in the market that no one sees. It’s like the dark matter of publishing.”
So the stigma may be gone for indie authors, but now it seems we might have to go through the same situation all over again with AI tools. Because the situation is similar.
There ARE a load of crap self-published books, as there are a load of crap traditionally published books, and as per Sturgeon’s Law, 90% of everything is crap.
Have you changed your mind about self-publishing in the last decade? Are there other technologies you’ve changed your mind about?
So yes, there will be a ton of crap AI-generated books.
But readers aren’t stupid
As JA Konrath said in his original article,
“Readers don't care if some moron uploads his ten-years-in-the-making opus “Me and My Boogers: A Love Story.” They’ll be able to avoid it just by looking at the crummy cover art, the poor description, and the handful of one star reviews.”
The same is true of the masses of AI books generated in one second or less which are flooding the store.
As with every other flood of crap content, Amazon/Google/Meta, etc will crack down and those books will be culled, the ‘authors’ penalized, etc.
Inevitably, some ‘real’ authors will get crushed by the hammer, but they will appeal and be reinstated.
There will also be regulation, safety guidelines, legal shifts, and guardrails added to the technology in the coming months and years as things develop. For example, we had Microsoft Bing’s Sidney for just a few weird days, and now they have added guard-rails to stop it getting too creative. [The Verge]
Like fire, electricity, and the internet in general, ‘AI’ is both a tool and a weapon.
It is our job to engage and help shape it for the better — because it is not going away and the usage and applications of AI will only increase and accelerate.
Compare the internet of 2003 with the internet of 2023. The impact of AI will be much greater than this. Read AI 2041: Ten Visions For Our Future – Kai-Fu Lee and Chen Qiufan for a glimpse of the possibilities.
So what about Clarkesworld and other smaller companies (agents/publishers/journals etc)?
Every company and publication will need to apply filters of some kind.
I am a one-person business, and I already use a premium paid spam filter and a premium paid Contact form on my blog, because I get hit by hundreds of spam comments and messages per day. These types of filter products will emerge for spam AI content and that will help somewhat.
There may also need to be a submission fee, even just 1c/1p, to submit to a magazine, or an award, to join an organization, or even to publish on Amazon. Some say that would prevent marginalized writers submitting, but there could also be a fund set up that people could apply for so that isn’t an issue.
Some say that’s against the spirit of these free submissions but times have changed, technology has changed, and business practices need to change, too.
Back to JA Konrath’s blog post, he ends with,
“If you're really worried about readers being subjected to crap, here's what you can do: DON'T WRITE CRAP.”
So now we come to the authors who love to write and create and who want to reach readers — and who want to use technology to help them do that.
Use AI tools responsibly to create great art and run a better creative business
There are ‘real’ human authors right now using AI tools to create better books and to run a better business. I am one of them, which is not a surprise to regular listeners as I have been talking about AI since 2016, and I wrote a book about the potential impact in 2020.
Usage of AI tools is a continuum and we all sit at different places along the spectrum. We are already a diverse group and each person is using the tools in a different way.
Some only use AI tools for generating sales descriptions or ad copy. Others are using it to generate finished words and images, then editing and shaping those according to their creative direction, as Tim Boucher does in the following interview.
I have been writing for publication since 2005. I first self-published in 2008, and every year since, I have learned about and used new tools that help me improve my craft, write better books and run a better creative business.
I embraced blogging and selling ‘ebooks’ (PDFs back then) before the international Kindle and the iPhone made mobile and e-reading popular. I started podcasting in 2009, way before it went mainstream. I jumped on YouTube and Twitter, and used print on demand for print books, and later used paid ads, and then tools like Vellum and ProWritingAid as they emerged — and many more.
The tools I use now in 2023 are almost completely different to the ones I used in 2005 when I started out — and many of the tools I use now are AI-assisted.
If you use Grammarly or ProWritingAid, or you use Amazon, Meta (Facebook), TikTok, Twitter, or Google for social media, search, publishing or shopping, you are AI-assisted. In the next few months, if you use Microsoft Word, you will be AI-assisted [The Verge].
So please, let’s distinguish between humans who will spam and scam and plagiarize and want to get rick quick, and those of us who are genuine writers who use AI tools as a collaborator/brainstorming partner/marketing assistant etc to help us create what we want to, in our voice, with our creative direction, and to reach readers with our books.
I remember being attacked for my publishing choices back in the ‘tsunami of crap’ days, and this feels like that time all over again.
The fear, vitriol and personal attacks make me just as upset, and make me want to stop talking about this stuff in public because I am just another human, wanting to make my art, wanting to reach readers, and wanting to make a creative living.
But back then, I felt so excited and energized and empowered by my choice to self-publish that I wanted to share my lessons learned with others.
And I feel the same way now.
In fact, I am more excited about the years ahead. I am having so much fun using the tools, and I am filling my journals with ideas sparked by these emerging technologies.
I love this stuff! I’ve been saying for years that this is the most exciting time to be an author, and once again, I stand behind that. What a brilliant time to be a creator!
I hope it doesn’t take a decade for this to shake out, but in the meantime — as I did 15 years ago — I will keep writing the best books I can for my readers, and continue to make a living as a full-time author entrepreneur, and I will use tools to help me do that.
We are in the very early days of “AI in everything,” and as such, it is a chaotic time.
But these tools are not going away, so you have to decide how much you want to engage.
Personally, I try to come at everything with an attitude of curiosity and playfulness. I spend a lot of time laughing and giggling with ChatGPT in particular, it’s like my inner 2-year-old come to life which is a lot of fun! You must find your own way.
Here are more resources to explore further:
- Ethical guidelines for AI usage from the Alliance of Independent Authors
- My backlist episodes on AI and the future of publishing
- Book: Artificial Intelligence, Blockchain, and Virtual Worlds: The Impact of Converging Technologies on Authors and the Publishing Industry
- The AI-Assisted Author course (includes aspects of ethics, copyright, bias etc)
I haven’t covered the copyright side here, as I covered some of the current legal cases in the intro to episode 670, my last AI show on AI-generated art with Oliver Altair, and I have an intellectual property lawyer coming on the show soon to talk more specifically about that.
Thanks again to my patrons who support episodes like this, and if you have questions to ask me about this or anything else, please join me at patreon.com/thecreativepenn or if you found this useful, you can Buy Me A Coffee.
Right, let’s get into the interview with Tim and continue the challenging conversation!
Transcript of the interview with Tim Boucher
Tim Boucher is a hyperrealist AI artist and writer specializing in questionable alternative realities. He's worked professionally in content moderation policy, and counter-disinformation efforts on behalf of a major web platform, a blockchain protocol. And he has advised nonprofits and governments on related issues. Welcome to the show, Tim.
Thank you. Thanks for having me.
Oh, I'm excited to talk to you. but first up,
Tell us a bit more about you and how you got into writing and publishing.
Sure, well, I'd like to say that I'm living the American dream of becoming a Canadian. So I was born in the United States, but I've been up here now since 2011. And now I'm a dual citizen.
I've always been interested in writing, I think the more recent place that I started to really do it again, a lot was in 2020, just before the pandemic, I started working on a novel called The Lost Direction. And that got me really excited again, about this idea that I could just be a writer. I could try to do that and try to make that work. It's not yet my full-time thing, of course, and that's the genesis of a lot of this other work that has come out after.
Tell us more about the lost Books project, What is it? And why did you create it?
Well, it's really just kind of my personal imprint that I use for some print books and a bunch of ebooks. but I've looked around a lot at publishing and professionally in self-publishing, and I like to kind of do things the way that I want to do them and not have to compromise about the direction of the work.
So making my own kind of imprint or press made more sense to me and then trying to pursue publishing through another avenue.
Lost Books, it's primarily right now, my current project is doing AI Assisted writing. but that's not the only kind of writing that I've done. As I said, the original book that I started with was all just normal manual writing, I don't know what we call that now, just regular writing, but
I like that ‘manual writing,' but that sounds like handwriting, so you would just be just a normal writer with no AI assistance.
Right, because in 2020, or whatever, this was in its infancy, and arguably it still is. but it was kind of only over the last six months or seven months that I got really into the AI side of things.
The tools started to improve, they started to be more available. And there started to be image generation tools that were really exciting, like DALL-E and Stable Diffusion.
And I've always been an artist as well as a writer, so having the opportunity to mix images and texts, I think, finally gave me the ability to produce a lot of things quickly that I've had, that I'd had stored up for a long time.
Because, like writing, to draw things by hand or to paint or even do things digitally, it's really time-consuming. So this has been a way for me to express all of these things that have been locked up for a long time.
What is the genre of Lost Books and also is The Last Direction different to these other books you've produced?
The Last Direction is pretty much just like a straight-up epic fantasy. And then there were some follow-ons to that one of which was called The Quatria Conspiracy. So that's sort of a, a very tongue-in-cheek take on the world of the original book, The Last Direction, proposing that it's not just a fictional story, proposing that there was a hypothetical ancient civilization called Quatria, that the whole thing is based on and that it's been covered up.
So from that I kind of had this idea of pursuing the presentation of my fantasy world, within the context of conspiracy fiction, it's a kind of switching it a bit to a pulp science fiction world where it's heavy on the world building.
It's heavy on alternate realities, sort of the uncanny valley between AI and humans. And this word hyperreality has come up a lot for me, where it's this sort of postmodern idea that fact and fiction get blended together. And I really took that idea and run with it. And AI tools have been really helpful in that regard. Yeah,
So when you first pitched me, I went to your store, and I was like, ‘whoa, this guy is creating AI generated books that are full of conspiracy theories.' And it made me worried because one of the problems that we're now seeing with these AI tools is they are not facts, that they make stuff up and for us as fiction writers, this is brilliant, but also, you've worked in misinformation.
Of course, someone's conspiracy theory is someone else's truth. Right? So how do you feel about this line between misinformation, disinformation, and conspiracy theory?
Yeah, I mean, it's one of the things that I want to talk about with my writing, one of the conversations that I want to generate. And I've had a little bit of success with this.
One of the first AI books that I put out was called Mysterious Antarctica, and it again, it follows this theory that there was this lost civilization and that they were in Antarctica, and that I used DALL-E to create images that were supposedly like people discovering artifacts, buildings and stuff in Antarctica in the 1950s.
From that, I got contacted for a couple of somewhat big fact checkers, because other people took my content out of its original context, and repurposed it on Facebook, TikTok, YouTube, Twitter, all over the place, posting it as though it's a real historical thing.
And I got fact check requests from Reuters, and from France 24. There was another one. So it showed me that the way that these things travel, you might start with one intent, but then other people sort of grab on to it, and they kind of twist it, and it changes.
But this is the nature of stories and storytelling in general. And something I saw working in disinformation is these alternative models of how stories are told, and stories are transmitted from person to person. And they get distorted and changed over time.
This, to me, seems like a very kind of modern take on folklore and mythology. And there are good things and bad things about that. There are risks when you take a chance like that as an artist and create something and present it in a certain way that people won't take it in the way that you meant it that they might distort it and look at it in another way, but I tried to mitigate that somewhat by I tried to only focus on sort of, I want to say fun conspiracies, things that are low harm potential.
I don't want to do anything that's against people or against groups, or that's too controversial or creepy. I tried to keep things that are more on the light side, like it would be fun if there was a lost civilization discovered in Antarctica.
What I've seen is that people are excited about this kind of way of looking at the world that it's like, what's maybe possible, you know, what's maybe true in opening up that kind of speculation, and then also being able to talk freely about like.
The problems around these technologies, like you said, they hallucinate facts, and they just say things that are wrong, or that are offensive, and that there needs to be that human layer on top of it, too. And that's one of the things that I'm happy about having these organizations ask me for a fact check. I want to be able to say What I'm doing and why and start these conversations.
Just to explain to the listeners, I said no to you, I don't want to talk about misinformation. And then you replied and kind of re-educated me on this. And actually, and now I agree with you.
And we're having this conversation because actually, we want to tell people, that there there is this myth of misinformation and what I mean, what's crazy, is it's not just the machine, it's also the humans.
Humans are spreading misinformation. Humans seem to love these stories. I mean, when you talked about Antarctica, I mean, I've read HP Lovecraft At the Mountains of Madness, and ancient civilizations, these stories, like you mentioned, myths, these are kind of part of our human mythos, and so they spread.
To me, the biggest thing that we should be doing is educating people about how to discover what is really true.
And we have to behave in a different way. We can't just assume that what we search for, or what we create, or what we use with these tools is correct. And so it's almost like we need to educate people to go deeper than we're used to.
And even with, like with Google. Google's going to implement its Lambda model. But perhaps we shouldn't be trusting Google anyway. I mean, I think we've realized that over the years, so I totally agree with you about this education of everyone about what we can trust and what we can't — and that includes humans, right?
People criticize AI technologies, because they can be “confidently wrong,” but that's exactly what humans can do, too.
So the same problems that humans create are the same ones we're seeing in artificial intelligence, but they're amplified because the technologies can scale and that's where risk starts to set in when something can be rapidly created and distributed and millions of people can see it in a short time.
So what did Reuters say when you replied, look, this is fiction?
I mean, they just published my comments for the most part, and there wasn't a lot back and forth on that particular point. It's interesting, I think there's an issue with the fact checks that if you're somebody that you've gone really deep into some of this alternative conspiracy stuff, sometimes a fact check has the opposite of its intended effect.
And it's something I saw just in the world of moderation. And the world of working with nonprofits and stuff is that even a well-intentioned fact check can sometimes backfire. And then some people can take that as proof of a cover-up.
So it's hard to communicate that thing. What you have to communicate instead is the tools for people to find the answer and be able to trust that they're going to try to apply them.
And that's kind of what I'm hoping to do is to be able to provoke that reaction, like, I see when people take my images and copy them and other places. And then people will come in and say, ‘Oh, that's fake. And here's why.'
I liked that people are coming in, and they're just applying their critical thinking on the spot, sharing it with other people, and starting those conversations, because maybe those people wouldn't be reached by a conventional fact check. but having this provocation helps to inoculate a bit against, I hope, some of the worst elements of that stuff.
There's lots to unpack there. I did want to come back to one of your blog posts where you were talking about creating these books, you say we invert, basically, you're “embracing the flaws of AI, writing the poor coherence, the inability to track narrative arcs, the tendency to invent facts, we want to recognize these artifacts as a feature, not a bug,” [Tim's blog]. I really, I really love that.
And I say that to people about AI audio. We're not trying to make it human. Once it is so like humans, we can't distinguish, that's probably more of a worry.
Talk more about the flaws of AI. Do you think they will disappear?
I mean, I think they're going to disappear. What we're seeing now is like the 1980s or 1990s of computing that, like we'd have these like primitive graphics, and that it would take a super long time to download over your connection, your modem.
And I think coming generations of technology, it's going to just go faster and faster. but to me, like, one of the things that is sort of the ideal purpose of an artist is to interrogate the technologies that they use to create and what I mean by that is that you don't just passively use them, you actively use them, you question them, you try to push them to the edges and to the limits.
And you document that process of here's what I found when I went out exploring with this technology, and some of its good and some of its bad. And here are the questions that it makes me want to ask it to find the answers to. And I think it's really important for artists to get involved in these kinds of new technologies, because there's a real risk of having a preponderance of only the engineering mindset.
Right, you know, I'm someone that I've worked a lot in technology and platforms and everything. but when I look at job listings, it's always for engineers. And it's like, I get it, that's the bread and butter of these technologies. but it can lead to lopsided ways of thinking about the human impacts of technology.
So I really encourage like artists and writers, anybody creative to get into these tools and become experts in them, like, take them apart, bit by bit and make your own tools make your own variations. Because when you become an expert, you can have a seat at the table in the conversation and the development of the technologies and you can help to steer it in a good direction. When you become an expert. And you have this firsthand experience, it's much more powerful than just sort of having a theoretical knowledge of it.
It's funny. I mean, obviously, I've been talking about this for years now. And I'm not I mean, I'm reasonably au-fait with technical things, but I'm not a programmer, and I'm not an engineer, for sure, but my feeling on this now is that let's assume that AI is going to get better and better — barring some big disasters, things will get better and better.
First of all, I agree with you, we should engage with these things, but equally, the AI tools ingest human artistic work so it will learn.
Someone emailed me yesterday, and they're like, Oh, I asked ChatGPT to roleplay as Joanna Penn and asked it to give me a pep talk about my writing.
And I said, ‘that's fantastic, send me the screenshots' and it was like ChatGPT was me. I love that. I have no issue with that at all.
But it was funny, because I kind of think, well, it's ingested all of me and I'm pretty positive. I mean, some of my fiction stories can be quite dark, but I have to believe that the machine will ingest all the happy things as well as all the sad things and all the beautiful art as well as all the terrible art and that it hopefully will come up on the side of beauty and goodness and all those things.
Although, I mean, there's always a dark side right, but part of me thinks that again, we're in such early days that it is only the technical people who are trying it, and a very few artists and writers and whatever, but that very soon this will be in everything, and people will be engaging with it, I guess in a more natural way, rather than trying to talk to the technical people.
I still come across these barriers sometimes that I'm not an engineer, but I've gotten fairly technically savvy about certain things. but there's kind of like levels of involvement in some of these things like that I still haven't crossed, like, when people link to like a Google collab notebook, I close out the tab, like, I don't want to get to the point of having to become a programmer to do the next level of stuff, but I think there are people that that's their skill set. And that's their interest who might not be programmers, but they might be able to make a real impact on that. So I think it's worth it for them.
There's a story that I like about the actor and director Buster Keaton when he first was presented with a movie camera. He took it home, or he took it to the hotel or whatever. And he took it apart piece by piece and figured out how everything worked, and then reassembled it and then kind of went from there.
So I think that's a really great way to approach tools and just like openly trying to discover them and experiment. And there's room for reluctance, because I understand that there are issues that need to be resolved. but don't let reluctance stop you from learning, from becoming educated, and gaining experience.
I totally agree.
I get a lot of comments, obviously, and emails and things, and I'm absolutely fine if someone has a reasonably educated view that is different from mine.
What I object to is someone who has not even looked at any of these tools, like someone who's not even gone to ChatGPT and tried anything, and then is just saying how awful everything is.
So just coming back to you being very open, and you have a blog where you share a lot of your process as well around this. And I kind of dug into that.
And I'm the same. I just had a blog post about my AI-assisted process for With A Demon's Eye, and I just list everything I used in terms of AI from images to text.
I was part of the ethical AI guidelines with the Alliance of Independent Authors. And one of the things that we're recommending is labeling.
So I label, my AI-narrated audiobooks have a little badge on them. I have this statement of AI usage in my books. I've been doing that. And what's so interesting is I've been criticized for using AI, and then I've been criticized for telling people that I'm using AI.
Right? So it's hard to win.
Yeah, it is hard to win. One cannot win.
But personally, I'm staying on the side of labeling even just to be aware of myself, like in 10 years' time, will I look back and say, well, it's interesting that I was labeling things, because clearly at the time, it was hard to know what was the right thing to do.
What are your thoughts on labeling and ethical and responsible usage of AI?
I think labeling is a really good and positive thing because it gives people an idea of what to expect.
If you're writing in a genre, you want to also label what your genre is, because people might not like that genre. So I think of it as a sort of a subcategory of whatever your genre is already.
But it's tricky because when people ask for labeling of content, they don't always know what it is that their objective is.
If you find out that something is AI-assisted or AI-generated, what does that tell you? It says different things to different people. And it's not necessarily clear what we should expect from that.
And it's tricky, too, because a lot of AI writing right now is hybrid-human and AI collaboration more than just purely AI-generated.
At a micro level, it becomes really difficult to say like, well, this part of it was written by AI, this part was written by me like how do we do that? Technically, as writers or within products that help us as writers that use AI?
So there are a lot of unanswered questions there that I'm interested in because I like tech products, and I like to try to figure them out. One thing I've thought about a lot is could we have sort of a simple markup or markdown way to within a text label, which parts are the AI-generated parts?
One of the things I've landed on is there's a historical mark of punctuation or the editor mark, I guess more called the obelus I'm not even sure how you pronounce it, but it's shaped in different forms. It's shaped like a division sign. Sometimes it's shaped more like a percent sign.
And I think what it was originally used for was in analysis of Homeric texts that people would mark on the side of the manuscript or whatever using this symbol, whether or not if they thought a part of the text was maybe invented that it wasn't actually true to Homer's original tale.
So there is a historical precedent for taking texts and marking them up to say this part is questionable, this part, we're not sure of the origin of it. So I think there's something there that can be explored of, of how can we do that and apply that at a really micro level within texts at an in line.
So I think that things like that, writers who have their own perspective and who have historical knowledge of analysis of texts, they're going to have a lot to contribute that an engineer might not know about. So that's another reason for people to get involved.
I think some of the other ethical issues I tend to think of, partly because of my professional background is in risk and analyzing risk, and then figuring out what do we need to do to reduce or to eliminate risks?
So I tend to think of things like internet technologies in terms of harms. Is there a harm to specific a specific person or to a specific group of people? Can you identify who they would be? And what would be the harm?
How harmful might it be? What is the likelihood? And things like that?
And then that gives me a more concrete framework from which to decide about? Who is impacted? And how much and why and is it really something that I want to engage in or not?
So I think something I'm seeing from artists and from writers who are hesitant to get into AI stuff is that I think people are waiting for ethical issues to be worked out over time. And I think that's a legitimate position. but I also think that if you're waiting for things to become perfect, and for all the possible problems to go away, you're gonna be waiting a really long time, and you're gonna miss out on being part of the conversation to develop it in a good and ethical direction.
For sure. I mean, on that, the ethical things, I mean, we even right now, as we record this, there's been all these articles about Roald Dahl, the children's author, I don't know if you've seen these about how his children's books are being rewritten to remove or anything considered offensive by a modern sensitivity reader. [The Guardian.]
And I grew up reading Roald Dahl. I'm in my late 40s now, and I still remember them very well. And they were brilliantly offensive, and I loved that as a kid, you know? And so there's a lot of hand-wringing over this.
Every generation has its own different decision on what is ethical and what is right.
So as you said, it will never stop. but I want to come back to that mark up in the text. That's definitely not what I meant by labeling, because personally, you mentioned hybrid writing earlier — that's how I'm using AI. It's in a very hybrid way.
There is no way at the moment with my writing process, I would label like, oh, this particular word is from AI.”
Tim Boucher: Oh, no, that would be too much.
Joanna Penn: Some people use it a lot more but what I also think is we will edit, so it is much easier to label an AI editor, an AI narrated audiobook, because that is it's final output. Like my book cover for With A Demon's Eye, that image is generated by Midjourney.
So that's easy, but I think text is hard.
Then coming back on harm. What is the harm? So let's come back to someone using ChatGPT to generate texts in the ‘voice' of Joanna Penn. Some people would not find that acceptable. Some people would say you should not be using my name in a prompt.
I'm delighted, but for some people that will indicate harm. You can go on to ChatGPT, you can say “write a blog post in the style of Joanna Penn about seven tips for first-time authors.”
Like literally go and do that now, and you could publish it under your name. Now, does that harm me? Where is the harm there? What do you think?
I mean, I think that has to be on an individual basis for creators because it's going to be a different limit for each person. I think there are risks there around the right of publicity of using someone's name or their likeness.
Or some people argue that style should be covered under that if it isn't already. I'm not sure exactly the technicalities, and then there's also the risk of impersonation, I guess that someone could use your name and then post something and then make it look like you posted something that it wasn't you I think it's just got to be figured out for each person.
But copyright law is not per person. Copyright law is copyright law.
So basically, the biggest question that is outstanding at the moment is are these bots trained on copyright data? And clearly, in some cases they are.
Copyright protects the expression of the idea, not the idea itself, so style is not covered. Right?
People already every day, take my blog posts and repost them on their own websites, even though that's a breach of copyright. So this is difficult. I mean, coming back on ethics:
One of my ethical rules is: do not use someone else's name in a prompt whether that's art or text or whatever.
Me too. I totally agree on that.
I don't want to create work in the style of someone else.
I want to create work in my own style or find a new style. I don't want to just go into Stable Diffusion, or DALL-E or Midjourney, or whatever, and say, make something like this person.
What's the next level? I always want to know, like, where it's going, what's gonna happen next. One thing I've seen, that's, I think, a bit problematic in some of the products that are, like third-party user interfaces.
For some of these tools, if you apply a filter and some of those products, it will automatically use an artist's name, and it might not be super apparent that's being added to the prompt, but it is. And I think that's kind of like a fishy area that I don't really like. I would rather just be more in control of the prompt itself and not try to imitate someone else and try to take it in a new direction. So I definitely agree with you.
So even though I said I'm not bothered about it, in general, it is not something that I think we should be doing. Because as soon as you use someone's name in a prompt, you know that machine is calling up whatever it has around that person, whether it as you said, whether it's visual art, whether it's music, whether it's written work.
If the work is in the public domain, then I don't have a problem. So for example, I have said to ChatGPT, rewrite this in the style of HP Lovecraft.
But then taking someone like Lovecraft, well known as racist and misogynist, all these awful things, but his writing was amazing, well, then obviously, I'm going to edit that and if anything comes up that is offensive in our generation, then I'm going to edit it.
So yeah, we're agreeing on this, right?
Tim Boucher: Yeah, absolutely.
Joanna Penn: Okay, so let's talk about some of the other things. So let's come back to your actual generation process.
What AI tools are you using for your books and covers?
I use a variety of tools. And for this reason, that also we were talking about, like the ability to track which parts were made by AI, it gets really difficult because I mix tools so much, that it's just hard to between like jumping between different tabs and stuff, as I go through a text that it's hard to figure out which part was written by which tool or which part was written by me. So I agree that there's a need to label it at a higher level.
But the tools that I use, in general. There's a website called TextSynth.com, that let you use a bunch of different open-source models like GPT J, and there's another one GPT Neo X and fair sec, I'm not sure how you pronounce that, but all of those are open-source text generation models, and they're accessible through that site.
And the way that that one works is you enter a bunch of sentences or, or however much text you have, and the AI tries to complete it. And then you can raise the temperature setting on this, which I like to do when I want to get like a curveball. And it starts to become much more weird and much more like AI feeling the text. So I'll use that sometimes.
I've been using verb.ai a lot, I think you use Sudowrite.
Yes, I have tried verb.ai. And as you're talking I mean, people are like, Oh, my goodness, I haven't heard of that one. There are literally hundreds of these.
There's a lot, yeah. But verb.AI I've liked because it's kind of straightforward. You just have your text area where you're writing and editing. And then they use forward slash commands, like you might have seen in Slack, or I guess probably Discord uses those too. You can use like a describe command, and then you can say ahat you want to describe or continue command, and then there's like a reword option and you can tell it like reword, but change it to be more whatever.
So I've used that one a lot. And I like it because according to the CEO who I spoke with, they use not only GPT 3, or maybe even 3.5, but they also combine it with some other sources, which I think are ai 21 and some proprietary texts that they use as corpus. So that one's nice because it's just like a simple editing environment. And it's relatively quick and it's I find that the text has a kind of unique and different flavor than what I've seen in some other tools.
I've also used a lot, character.ai, I really like them, you can go and within just like a few seconds you generate a character, describe what it is, and then you can converse with them. You can have it be a character in your book or your story, and then you could also have it be a historical figure or some secondary figure.
I've used that that site a lot to work on, like press releases, or to work on like more of the short expository writing that goes on around promoting a book, which I feel that I'm less good at or less confident in. And that does a really good job because like I can have a dialogue with this tool that I say what I want, and then like reconfigure it to ask me some questions. And it's not recreating everything from scratch. It's not coming up with the most creative things ever, but this act of like the dialogue, it lets you to get somewhere that you wouldn't have gotten if you hadn't engaged in that. So it's really good for that.
ChatGPT is similar in that way. It has kind of a different focus, though. It's almost to me not as much meant to be a chatbot. I don't go and ask ChatGPT what it thinks about things, you know?
It's helped me rewrite sales descriptions to make them more like a best-selling thriller.
And I think it does a great job at that kind of stuff. but I found that using ChatGPT for like, more of the creative writing side of stuff. It's a little bit flat sometimes.
Though again, we're speaking on the 20th of February 2023, and just before we got on the phone and article, I just saw an article [Silicon Republic] saying that they're going to in their advanced paid service for ChatGPT, they're going to allow you to configure what you want the service to do and change the guard rails.
I got on ChatGPT the very day they launched it, like that the morning I was on it. And what I was able to do around darker fiction stuff, like even guns, right? I write thrillers, there are guns, whatever people's ethics are, that is in the thriller genre.
And so I was like, wow, this is great. And then within a couple of days, you are not allowed to talk about guns, right? The service has got more and more restricted. Because they're having to shut down this and shut down that but then people are like, well, what if I have conservative politics? Or what if I am someone who wants to write about guns? Or what if I am someone who wants to write about Marxism?
What they're saying is, you'll be able to configure the guardrails (presumably within reason) in the future. So I'm hopeful that it will mean I can get back to what it was at the beginning. We shall see of course.
No, I agree about a lot of that. And I think the thing for me that I found the most annoying as a fiction writer is like, I'll be like, Okay, write a science fiction thing about this. And then it will be like, Oh, as a large language model, I'm not able to do that, blah, blah, blah.
You have to say, “Pretend you are an author, or we're role-playing where you're an author.” You have to like trick it.
And I totally understand OpenAI's perspective in that they've got now I think, like 100 million users or something. And it's really hard to satisfy the needs of that many people through a single tool with a sort of a single configuration. And I think we've seen how that plays out. Everybody wants something different from it. So I think that's going to be a good direction, we'll see how it gets deployed.
Going back to the tools.
I use DALL-E a lot for images, I think that their images have like a kind of a luminous quality.
To me, there's something about just like the colors on the light that I like, when you get into certain aspects of it. I've used Stable Diffusion quite a lot to use, specifically playgroundai.com because they let you generate a bunch of images for free or there are paid tiers.
And most of the time I start with my text first, I start with the title. I have sort of like a few different formats of titles that I use. I haven't mentioned this earlier in the podcast, but I am up to like 67 titles now that are AI-assisted.
How long have you been creating?
I started on August 4, 2022, so it's a rate of about 10 a month, a little bit less. I should stipulate that like, those are not full-length books, generally they're between 2000 and 5000 words and they include between 40 and 100 images.
Are you doing this all day, every day or have you also created some kind of program that is just doing it for you?
No, I'm just like a voluminous writer. I'm very productive. And I'm pretty efficient. And I've really nailed down a workflow that just lets me operate fast.
And to some degree, it's a formula. but a lot of genre writing becomes a formula. When you have a formula, there comes a lot of freedom, because you know what you need to hit in the overall structure in your writing, and then you're free to explore all this other stuff around that.
So having sort of a structure, a workflow, a formula that I can rely on and fall back on, has helped me to build exponentially what I would do just manually before.
What does your AI-assisted workflow look like?
But my workflow in general, I start with the title, and I generate maybe a sentence or two about what this volume is going to be about.
And then I'll just go and start writing it in verb.ai or another tool, and then I'll pop in between different tools and have things get expanded or changed and add another element.
And then once I've kind of reached all the stuff that I wanted to reach in that volume, I will go and I'll do my image generations. And then I save those locally.
And I put them into Adobe Lightroom, which is a really great product that I think few people know about fairly for this use for managing large sets of images and being able to kind of quickly go through and pare down out of all the image generations that I did into the golden set that's going to fit best with this title on this topic and the vibe, the mood and everything.
And then from there, I go into Vellum, which is my favorite ebook writing and publishing application by far. Really simple, really fast, very beautiful results, almost no problems ever with it. So I really love them.
And I'll put everything into short chapters, and then I mix the images into the text and I try to have sometimes the images are reflecting the contents of the text, and sometimes they're like really clashing about it like, it will have sort of a different visual story or like a complimentary visual story that sort of like highlights and expands on what's being presented in the text.
I should also probably note that structurally, ahat I'm doing as a writer is, it's really a lot of world-building and lore. I have like very few deep characterizations, you know, like, sometimes books, like there's just not even a character, it's more like an encyclopedia entry about this fictional thing.
And this is something that a rich precedent of that I think especially in like, mid-century sci-fi, of having like Dune, or the Foundation books that a chapter will start with like an excerpt from a fictional encyclopedia. So that's a structure that I play on a lot.
And it lets me incorporate, like I said before, these like flawed factual statements, or these like inconsistencies, and just present them as they are, and not try to dress them up.
I think if you're doing other kinds of writing, and you're using AI, you might end up being more frustrated than then my use, because it does have a hard time keeping track of characters and in keeping track of narrative arcs like tools like verb.ai and, to some extent, some other ones are, they're trying to solve that, but none of them have quite landed on how to do that.
They're getting there, though, like Sudowrite now is basically you keep a little side note with things, and then that gets incorporated into the next chapter.
But coming back on the speed. I can hear people who are listening, they're still on “67 books since August.” !!!!
Let's just be really brutal. People will say AI-assisted books are a ‘tsunami of crap.'
This used to be thrown at self-published authors back in the day. We were told that self-publishing is a tsunami of crap that will drown out real books, real authors. [JA Konrath, 2011.]
Now, I think this is like 2007 all over again, this is a tsunami of crap, the quality is bad. There'll be no real books, no real artists anymore, it will just be awful, and it will kill off real authors, etc, etc.
So what do you say about that? I mean, inevitably, you've talked about making 67 books. They could be the very top end of literature, but still the ability to make them that fast means we are going to have an explosion of content like we've never had before.
So what are your thoughts on this issue of quality and tsunami of crap? What can authors do to make good art and also to stand out in a very, very crowded market?
Yeah. I mean, I tend to follow this prediction by a journalist named Nina Schick. She has a generative AI newsletter on substack that's worth following too.
She predicts that within a couple of years, 90% of content on the internet will be aI generated. Honestly, I think it might be higher than that. Just because of the sheer fact that AI can do it much faster than people and even myself, like, I'm not using some program automated to automate the whole workflow. I'm just like using tools and then piecing things together.
And I think eventually people are going to automate the whole workflow. And it's going to be more of a single button press. And I think that's, in some ways, that's bad, but in some ways, like if the content that's generated is good, what's the major issue? But of course, not all the content that's generated like that is going to be good.
Well, the major issue is that we all find it hard enough to make a living selling books now when there are already something like 21 million books on the Kindle Store. How will it work when there are 21 billion?
Or how are people going to find my Shopify store?
So marketing and making money with your books is the problem.
But let's leave aside the quality discussion because that is in the eye of the beholder. And I totally agree with you, I think within a very short time, you will be able to press a button and output a thriller, or a romance according to whatever rules you like, but how are you thinking about this?
How are you building an audience? How are you marketing? How do you think we can stand out?
I mean, I think, like, writing manually in this mode, as we said, like, I don't think that's going go away. People want to do that, and people want to read that.
So I think there's room for everything in the future that I imagined my ideal vision that's like that, if people want that they can get that if they want things that are more AI-generated, or that are like custom-tailored to their specific interests or whatever, like, they'll be able to get that too.
And I think there are going to be issues about like, how do you position yourself within that market? And I think that's where becoming an expert in how the tools work is going to be to your advantage.
There's also an Alan Moore quote that I remember, I think he was talking about V for Vendetta, like his comic book, the original vision versus like, how it was transformed into a movie. And he said, look, the comic books still exists, it's always going to exist in that form. And the movie is this other thing. And people can take it in, in the way that they take it. but he didn't see that conflict between the movie replacing the comic book, because it's still there. It's still available.
But I agree that marketing books is one of the hardest things that I've ever done.
Because also, you don't sell on Amazon, do you?
No, I sell only on gumroad.com, which is probably weird to people.
But just briefly, Amazon has such a controlling influence sometimes in my life. I'm always buying things on Amazon, that I feel creatively, I don't want them to control my creative output. There's something creepy about that.
For me, if I'm wearing like Amazon sweatpants, and watching Amazon shows, and they own all my books, that feels kind of dystopian to me. There are so few corporations that that control so much of publishing, I think that's a really negative thing.
So I think that being more DIY, even than Amazon, and carving out these alternative ways to sell it, it can be one way to differentiate yourself too, because you can't go on Amazon to get my books, you have to go through Gumroad.
And when you do, I know exactly how much Gumroad is going to take, they're going to take 10%. And I also know that if you buy a book, I'll get your email address, like I'm not. So far I'm not using those and I'm not doing a newsletter or anything like that to promote. but I can see like, if someone bought a book, I can see like how many books they bought, like I have a lot of people who they come back after buying one book, and they'll buy like six more, they'll buy 10 more, I have one person who's bought like over 30 titles out of 67.
Was that Mr Tinfoil hat?!
Well, you know, their interpretation is, I'm sure it's going to be different than mine, but I don't want to also enforce a single interpretation of the work either. But I think there's there's a benefit to having greater control over not just the tools of production, but also the distribution.
I mean, I obviously talked about selling direct all the time, but still, the question remains, how do you get people to find your Gumroad store?
Right? So because my work is like a blend of fact and fiction that rides on this edge of science fiction and conspiracy, like, I will post things to want to Reddit or my Medium accounts. Those are really the only two social media sites that I use anymore.
And from there, people go and they read the pitch page, and if they're interested, they buy it.
I'm doing pretty well compared to how I was doing when I was just trying to promote an epic fantasy novel. Promoting an epic fantasy novel is really hard. And if I had known that before, maybe I would have been more disillusioned before going in but it's part of why I've modulated my approach.
It's part of why I got into this idea of the hyperreal storytelling devices and just framing things in different ways and letting people make their own decisions about what's real and what's not.
One of the things that's ironic is that my original fantasy book, The Last Direction, it's sold far fewer than any of the other follow-ups about like the conspiracy framing of it of the Quatria conspiracy, or the Mysterious Antarctica book like that has sold so many more.
But Mysterious Antarctica, or Mysterious Egypt or Mysterious whatever, that's the kind of thing I buy for my research, It's fun, even though it's fake, and I read books for ideas.
We could talk forever, but I feel like we're almost out of time. Before you go, we've talked about a lot of things that are difficult, a lot of things that are challenging for people, but where do you think things are going?
Because I really do think we're in the super, super early days of all this.
It feels like 2007, this kind of iPhone moment when we were like, What do we do with that? Who needs a phone like that? And this is like now. So what's it going to be like in the next 15 years? What are you excited about? Obviously, not 15 years away, but what are you excited about? What do you think's going to be happening?
I mean, one of the things that I'm most excited about is like, Okay, I want to reach 100 volumes, and I'm like, feverishly trying to work towards 100. I calculated that it's going to be like maybe another three and a half months if I'm lucky and I can keep up the speed.
What I'm excited about is, is eventually having another AI layer that can come in, it can ingest all my books and it can make like a map of all of my universe, or multiverse, or whatever.
And it can say these are all the relations between all the entities and the characters and the places. And here's a timeline of events, and then it can reference the different sources. And then you could even like highlight, like, okay, there's a conflict between how this is described here and how this is described there.
And I think like, once I can reach that level where I can put the work in there and then get these like next-level interpretations and representations, there's going to be like a whole new layer of storytelling, that's going to be really exciting.
And it just seems like an inevitability that that will happen. And I've already seen, there's a tool called file chat, that you can upload your documents to it. And then you can sort of like conversationally query the contents of the documents. So I think that's not going to be a super long way away, kind of what I'm describing.
I agree. And I mean, we talked about the problems of discoverability, but actually an AI over layer that can go actually find something for me, that can actually find me fiction that I would love regardless of where that might be, I think that'd be brilliant.
Because it feels like at this point in time, the Amazon algorithm is broken, or is gamed by ads or whatever. And I feel as a reader what's so crazy as a reader for the last couple of years, really, I've found books in new ways. I've gone back to bestseller lists, or I've gone back to physical bookstores, because I feel like the ad model has changed things and that the recommendations I get are not good enough. So I want a better AI to recommend me books.
I have the same problem with music on Spotify that like, because I listened to something a few times. Now Spotify is convinced forever that this is my favorite thing. And I only want to listen to that from now on, and it gets very distorted the recommendations. And it makes me feel that like, now I don't even like the things that I like before because like Spotify is just forcing me on it.
And I contrast that to a decade or two decades ago when I was younger. And like when my main source of new music and movies and books and stuff was from my friends, like people who I hung out with in real life, that would be like, Oh, you're gonna love this, check this out.
And to me, that human element is never gonna go away, but I think like you said that over time as the AI is approved, we'll get better recommendations and better ability to find the things that we're going to love and that are going to change your life.
And I think one solution for writers is to be super, extremely niche. Because like, as those recommendations improve, it's going to be easier for people who follow that very super specific niche to find you.
Yes, so a combination of exciting times and difficult times.
I said to Jonathan, my husband the other day, I just love being alive right now. It's so exciting. Using the surfing the wave analogy. We're kind of wobbling on these boards. We don't know what's under the water, but it's really exciting at the same time. You feel excited and terrified, but yeah, this is amazing. Let's go!
Yeah, yeah, let's go!
So where can people find you and your books online?
Yeah, so the books themselves are at Lostbooks.gumroad.com or you can go to LostBooks.ca. Also, I run a personal blog, and I've rediscovered my love for just being able to be true to my vision of something and not having to worry about how it appears on a platform. I just get to write it myself. So if you want to see that stuff, you can go to TimBoucher.ca. And I talk about process and the technologies of AI and how artists need to have a strong role. So if you're interested in that stuff, go there.
Fantastic. well, thanks so much for your time, Tim. That was great.
Seamus Anthony says
I’m sorry that some are giving you a hard time Joanna. Whatever anyone’s opinion about AI, it’s no excuse to carry on like that. They need to take a deep breath and deal with their fear. Please ignore them and continue to bring the awesome for your fans.
Joanna Penn says
Thanks for your support. I try to ignore, but sometimes the comments cut a little deep!
Probably because I care about my art and my books and my audience, so when some accuse me of the opposite, I want to put them straight.
Anitha Krishnan says
Hi Joanna, it was after listening to your podcast that I tried out ChatGPT and it’s been such a great tool to play around with. Thank you for approaching AI with curiosity and a sense of empowerment; you have inspired me to adopt the same attitudes and consequently, far from feeling threatened, it’s been exciting to witness all these changes happening. Listening to you over the years has really helped me develop a positive mindset in this indie publishing journey. Thank you!
Joanna Penn says
Thanks, Anitha, and I am so glad you’re finding the tools useful 🙂
Marc Holmes says
This was a great listen. I’m taking on the conversation while sketching designs for a graphic novel – and thinking quite a bit about AI assistance. An obvious use, which I have not seen yet, is AI *coloring* (or painting) of comic art. I am aware AI can do the whole thing, but, as with your guest, I’d like to see what my style would look like in collaboration. Will see how it goes.
Joanna Penn says
It’s still early days, but I have seen some ‘half-colored’ sketches on Midjourney, so presumably it’s starting to happen … Fun times!
Regina Clarke says
Interestingly, Tim says “I’m always buying things on Amazon, that I feel creatively, I don’t want them to control my creative output. There’s something creepy about that.” But he has no problem letting AI control his output to some degree, whereas the idea of that is a bit creepy to me. The author’s voice is no longer their own, bad or good, when combined with AI. It makes me think of the Russian literature under Stalin called socialist realism, where what was created was not the author’s work solely, but culled by the structure of the politics at the time. AI is to me such a massive opening into something not wholly the author’s imagination. Having read literature all my life, it is very hard to accept the idea that AI’s distortions of an original author’s theme are seen as legitimate. That said, I can see why it could draw a crowd–it is a whole lot easier to generate AI in seconds than create a book over the course of months or years. But I bet Ray Bradbury would have more to say on this, and say it better than I.
Joanna Penn says
I use the tools more like a co-writer and the ideas thrown up spark new ideas in my imagination. There is no distortion — it’s amplification.
I find the tools fun and inspiring to use and those I see using them are loving creating more than ever.
Regina Clarke says
My favorite team in scifi, the cast of Stargate SG-1, gathered together and did a reading at a conference last year that was an AI-created script for a hypothetical episode of their show (a show from 1998 to 2008 which I dearly wish had gone on forever). It had coherence, adventure, identifying characters accurately (bit creepy, that) and so forth. It was a fun session and they were open to the idea, but it made a point I feel is true–there was no soul in the AI version. There was no overriding feeling of happiness listening to the episode–it had nothing to do with what made the show a success–the genuine camaraderie of the original characters. In essence, that is a spiritual aspect that a machine cannot duplicate or render. And in essence, that is what I feel AI hinders even in a collaborative way. I’ve used it, briefly, experimentally. I expect in generations to come the original voice of a writer may disappear entirely with AI–it’ll be like the writer of a movie script, where generally speaking everything the screenwriter writes is up for modification by everyone from the lead actor to the cameraman. And the original story is only what it was intended to be if the writer is either also the director or applauded by the director.
Tim’s argument, and yours, will no doubt win out. We will see a story half ours and half (or more) a machine’s as the status quo. It is simply too tempting. Or in the view of you and Tim, just a very legitimate form of creating literature. But when I got the message from Clarkesworld (where I have sent stories) that all submissions had been canceled because of the onslaught of AI, I was shocked, aghast. So I wrote to the editor of another very respected short story venue for writers in fantasy and scifi about whether they had also experienced the tsunami Clarkesworld had. And they said they had to look at things, because the increase in submissions since the New Year was overwhelming. Filtering out bot versions of stories is not easy. So instead of reading several pages of submission, they are considering reading just a paragraph or two, to manage the doubled volume, which of course means real authors (aka non-AI authors) will get far less of a viewing/hearing. All the major short story genre magazines are affected. (Not so much literary, perhaps because they are not usually given a wide audience in the first place.) I can see many short story magazines folding under the weight of this.
Anyway, you have written about this for a long time and I remember you advocated it even on one of your sessions with Orna a few years ago. It has touched a chord for you and you are in a position to make your advocacy heard now on a wide basis.
Alas, my voice is not likely to make a dent… 🙂 One thing that has occurred to me is that this may mean the demise of all scifi stories about robotic entities. Where’s the fun if it’s a robot writing it?? It is the voice of the original author I will miss the most, that inner, creative force seeded by life and spirit, unique to the writer. I wager the idea of what is unique about a writer’s vision will have to be redefined, too.
Thanks for replying. Given your passion for AI, I rather didn’t expect you to answer a message saying the opposite… 🙂 Decent of you.
Oh–I saw your AI-generated art for the next Arkane thriller recently–so I expect that book will have AI elements in the text, too. You are one of the first responders to this realm.
And on Tim again–when he cited all the ways our computers and companies already use AI–I have no objection to AI for ads and nonfiction how-to’s and all–creating that stuff is such a chore–and it is great how AI can simplify so many processes we already all use. No, no problem there. My objection to it–or better, my lack of feeling for AI when it comes to fiction comes out of my love of inner inspiration, and the wonder and wondrous feeling for what I have made evident on the page through my own heart and spirit. I just don’t buy that AI can offer that.
But again, alas, I have a very small audience at hand to tell of this… 🙂
Joanna Penn says
“It is the voice of the original author I will miss the most.”
Not everyone will use the tools, Regina, so don’t worry, you can write without AI, as will many others.
On the other hand, what about all the people whose voice you WILL be able to hear with the use of these tools? I am getting a lot of emails from people with chronic illness, disabilities, language issues, and other things that have stopped them creating what they want to in the past, and now they can. Their voices will finally be heard, and that is surely a good thing.