I love how you've embraced this and are (as always) so willing to share your experience and ideas in practical, encouraging ways. I attend a weekly meetup on generative AI, and a slightly revised version of quote I wrote down from one of the presenters a couple of weeks ago rings true here: "AI won't replace marketers. But marketers who use AI will replace those who don't." Swap marketers as you will… or ask ChatGPT to do it. ;-)
These are clever ideas! I definitely concur that the best use of chatGPT is for stuff like idea generation instead of actual writing. When you get it to come up with comp titles, how often does it hallucinate/lie and pitch you things that don’t exist or don’t fit? I’ve had it fabricate whole films and people before so I don’t trust it as a research tool without some hefty double checking lol
I love the way you use ChatGPT and Tarot cards for protagonist and antagonist motivations and story ideas. I hadn't thought of doing that, but it sounds like it could be a terrific tool to boost creativity when mine is low! So thanks for that inspiration.
I use it for non-fiction as you do - a starting point. For idea generation, but not to do the writing for me. Google and other platforms penalize sites for content that is AI made over human-produced content, so the lazy writers that think AI can do their jobs for them will be hurt in the long run. But as a tool and partner, it's proven to be worth the time it takes to use it responsibly, for sure. Idea generation can sometimes be just the kick in the pants we need as non-fiction writers. But yes, it's accuracy should always be checked for sure!
Great post, I look forward to any other insights and responsible uses you find for ChatGPT.
I like the idea of using it to come up with title ideas. I’ll have to try that!
Mostly, I’ve been using it to generate niche lists, like “common jobs in post-apocalyptic movies,” that help me think of a specific I need for the plot.
I had to try it. I pulled my Tarot deck out and shuffled, thinking carefully about the protagonist, antagonist, protagonist goals, resolution and conflict in turn of my story that is 1/3 already written. I already have very developed characters, conflict and all that.
The first paragraph in the returned results was SPOT ON.
"In this story, the Knight of Cups represents our protagonist, a romantic and emotional individual who seeks fulfillment in life. However, they are constantly challenged by the three of swords, the antagonist of the story. This card symbolizes heartbreak and betrayal, and our antagonist seeks to undermine the Knight's emotional stability."
The results on the conflict and protagonist goals was less than accurate, but the result on the story resolution and overall protagonist arc was closer to my vision for the story than I expected.
I admit, I'm one of the big nay-sayers about the use of Chat GPT or generative AI in any of my work. However, my genre is one of the ones targeted by scammers misusing the tool--I write SFF short stories that I submit to the traditional publishing markets, and had noticed before Neil Clarke wrote about the story submission swarm that the bigger short story markets had suddenly universally implemented "cooling off" periods between submission responses, anywhere from one to three weeks.
That immediately impacted me as I have stories making the rounds at these markets. Instead of processing submissions on a weekly basis, which is best practice, I'm now down to processing submissions on a monthly basis. I also am reconsidering whether I should remain in the short story market as a result. A submission swarm flood from AI-generated sources could shut down open windows at some markets. Coupled with Amazon dropping its magazine distribution program, that also means a more competitive, smaller marketplace for short work. I am NOT HAPPY about this at all.
I'm also seeing publishing scams emerging. High-priced marketing courses going for almost $1000, touting THEIR take on ChatGPT.
But here's the other piece. I do not use outside methods to generate story prompts, plot, develop characters, or the like. I've never had much patience with them and don't have issues with idea generation.
Knowing what I do about measurement of human intelligence (ten years as a special education teacher and caseworker where I worked a lot with cognitive measurements), I'm skeptical of the long-term use value of ChatGPT for anything other than routine drafting of the most mundane work. I already fight with the AI in the Office 365 version of Word because it seeks to homogenize my voice (today's plan is to figure out if I can, if not kill the damn thing, at least dismember it so it doesn't annoy the crap out of me). Every example I've encountered of ChatGPT in creative work demonstrates that reduction of individual voice.
The other problem is that it hallucinates. It fabricates. I don't trust it for those reasons alone, because it won't get rid of the need for extensive fact-checking. I also do not want to condone the training process because much of that training was done on work under copyright without the copyright holder's permission.
And, speaking from my background in cognitive measurement, albeit in applied cognitive measurement and not theoretical? It is still a mechanical intelligence, and not an example of artificial general intelligence. At the end of the day, it is still just a very powerful crystallized intelligence (i.e., it knows a lot but it is not capable of making human intuitive reasoning jumps, especially when presented with a novel situation) that operates on brute force manipulation of that knowledge.
I want to keep the human in my work. And that for me means that I don't use it for anything other than the occasional promotional practice, if that--because I also find that the marketing work that others find annoying and redundant also gives me insights into how a particular book works that will also help with future promotion. I am not looking forward to the next wave of repetitive, cloned promotions out there.
I had my whole worldview blasted open when I recently used ChatGPT to come up with a pair of brand new terms I'm going to introduce in an essay about ADD. I had to do the research, and give it the right prompts, but it came up with a solution it might have taken me weeks of pain and toil to get to.
The problem I have at the moment with generative AI is that even if you're an ethical user (and I 100% believe you are), there's a deep uncertainty about the technology itself. Is generative AI based on unethical tech at its core? If so, then I'm not sure how possible it is to be an ethical user.
That uncertainty is why I'm steering clear for now. Hopefully someone is working on interesting, generative AI that has a clearer and more honest relationship to its learning material.
I recreated your first Tarot GPT results with Microsoft Bing. This is what it came up with:
Sure! I’ll do my best to provide you with a comprehensive reading of your three tarot cards using the past-present-future spread.
The 2 of Swords card represents a decision that needs to be made, but you are feeling stuck and unable to make a choice. You may be feeling like you are in a stalemate or that you are at a crossroads in your life. This card suggests that you need to take some time to reflect on your situation and weigh your options before making a decision.
The Daughter of Cups card represents creativity, intuition, and sensitivity. You may be feeling inspired and ready to start new projects or ideas. This card suggests that you should embrace your creativity and intuition and let them guide you towards your goals.
The 6 of Pentacles card represents generosity, charity, and sharing wealth. This card suggests that you may be experiencing financial success or abundance in your life. You may be feeling generous and wanting to share your wealth with others.
When we look at these three cards together, we can see that they suggest that you are at a crossroads in your life where you need to make a decision (2 of Swords). However, you should trust your intuition and creativity (Daughter of Cups) when making this decision. The 6 of Pentacles suggests that if you make the right decision, you will experience financial success and abundance.
I hope this reading helps! Let me know if there’s anything else I can do for you.
Great article! Most articles I read about using ChatGPT come from copywriters and business writers. I'm happy to see the fiction and literary perspective.
Bit late this to chat, but to the point several people make about the issue about AI being about the use people make of it in the area of creativity, I'd like to submit an example. Realising as a unknown author, I was unlikely to be interviewed about my book, I got ChatGBT to do it. The result amazed me and might inspire others to think laterally about how to use the tool: https://open.substack.com/pub/invasionlovebirds/p/a-famous-influencer-interviews-the?r=cfm0p&utm_campaign=post&utm_medium=web
I love how you've embraced this and are (as always) so willing to share your experience and ideas in practical, encouraging ways. I attend a weekly meetup on generative AI, and a slightly revised version of quote I wrote down from one of the presenters a couple of weeks ago rings true here: "AI won't replace marketers. But marketers who use AI will replace those who don't." Swap marketers as you will… or ask ChatGPT to do it. ;-)
These are clever ideas! I definitely concur that the best use of chatGPT is for stuff like idea generation instead of actual writing. When you get it to come up with comp titles, how often does it hallucinate/lie and pitch you things that don’t exist or don’t fit? I’ve had it fabricate whole films and people before so I don’t trust it as a research tool without some hefty double checking lol
I love the way you use ChatGPT and Tarot cards for protagonist and antagonist motivations and story ideas. I hadn't thought of doing that, but it sounds like it could be a terrific tool to boost creativity when mine is low! So thanks for that inspiration.
I use it for non-fiction as you do - a starting point. For idea generation, but not to do the writing for me. Google and other platforms penalize sites for content that is AI made over human-produced content, so the lazy writers that think AI can do their jobs for them will be hurt in the long run. But as a tool and partner, it's proven to be worth the time it takes to use it responsibly, for sure. Idea generation can sometimes be just the kick in the pants we need as non-fiction writers. But yes, it's accuracy should always be checked for sure!
Great post, I look forward to any other insights and responsible uses you find for ChatGPT.
I like the idea of using it to come up with title ideas. I’ll have to try that!
Mostly, I’ve been using it to generate niche lists, like “common jobs in post-apocalyptic movies,” that help me think of a specific I need for the plot.
I had to try it. I pulled my Tarot deck out and shuffled, thinking carefully about the protagonist, antagonist, protagonist goals, resolution and conflict in turn of my story that is 1/3 already written. I already have very developed characters, conflict and all that.
The first paragraph in the returned results was SPOT ON.
"In this story, the Knight of Cups represents our protagonist, a romantic and emotional individual who seeks fulfillment in life. However, they are constantly challenged by the three of swords, the antagonist of the story. This card symbolizes heartbreak and betrayal, and our antagonist seeks to undermine the Knight's emotional stability."
The results on the conflict and protagonist goals was less than accurate, but the result on the story resolution and overall protagonist arc was closer to my vision for the story than I expected.
So for this use...I'm impressed!
So interesting thank you what a great post
I admit, I'm one of the big nay-sayers about the use of Chat GPT or generative AI in any of my work. However, my genre is one of the ones targeted by scammers misusing the tool--I write SFF short stories that I submit to the traditional publishing markets, and had noticed before Neil Clarke wrote about the story submission swarm that the bigger short story markets had suddenly universally implemented "cooling off" periods between submission responses, anywhere from one to three weeks.
That immediately impacted me as I have stories making the rounds at these markets. Instead of processing submissions on a weekly basis, which is best practice, I'm now down to processing submissions on a monthly basis. I also am reconsidering whether I should remain in the short story market as a result. A submission swarm flood from AI-generated sources could shut down open windows at some markets. Coupled with Amazon dropping its magazine distribution program, that also means a more competitive, smaller marketplace for short work. I am NOT HAPPY about this at all.
I'm also seeing publishing scams emerging. High-priced marketing courses going for almost $1000, touting THEIR take on ChatGPT.
But here's the other piece. I do not use outside methods to generate story prompts, plot, develop characters, or the like. I've never had much patience with them and don't have issues with idea generation.
Knowing what I do about measurement of human intelligence (ten years as a special education teacher and caseworker where I worked a lot with cognitive measurements), I'm skeptical of the long-term use value of ChatGPT for anything other than routine drafting of the most mundane work. I already fight with the AI in the Office 365 version of Word because it seeks to homogenize my voice (today's plan is to figure out if I can, if not kill the damn thing, at least dismember it so it doesn't annoy the crap out of me). Every example I've encountered of ChatGPT in creative work demonstrates that reduction of individual voice.
The other problem is that it hallucinates. It fabricates. I don't trust it for those reasons alone, because it won't get rid of the need for extensive fact-checking. I also do not want to condone the training process because much of that training was done on work under copyright without the copyright holder's permission.
And, speaking from my background in cognitive measurement, albeit in applied cognitive measurement and not theoretical? It is still a mechanical intelligence, and not an example of artificial general intelligence. At the end of the day, it is still just a very powerful crystallized intelligence (i.e., it knows a lot but it is not capable of making human intuitive reasoning jumps, especially when presented with a novel situation) that operates on brute force manipulation of that knowledge.
I want to keep the human in my work. And that for me means that I don't use it for anything other than the occasional promotional practice, if that--because I also find that the marketing work that others find annoying and redundant also gives me insights into how a particular book works that will also help with future promotion. I am not looking forward to the next wave of repetitive, cloned promotions out there.
Here's the meetup info, Jackie. It's an open group. Next one is tonight: https://www.meetup.com/ai-artists-salon/
I’ve used it for writing job descriptions and job adverts at work.
For my newsletter I often use it for the title and for generating a few ideas to get me off the blank page.
But...it’s so important to add a human touch. Your human tough.
It’s obvious when you don’t.
I had my whole worldview blasted open when I recently used ChatGPT to come up with a pair of brand new terms I'm going to introduce in an essay about ADD. I had to do the research, and give it the right prompts, but it came up with a solution it might have taken me weeks of pain and toil to get to.
Really interesting uses and experiments, Jackie!
The problem I have at the moment with generative AI is that even if you're an ethical user (and I 100% believe you are), there's a deep uncertainty about the technology itself. Is generative AI based on unethical tech at its core? If so, then I'm not sure how possible it is to be an ethical user.
That uncertainty is why I'm steering clear for now. Hopefully someone is working on interesting, generative AI that has a clearer and more honest relationship to its learning material.
Wow, that's interesting stuff. Thank you for sharing that. I wasn't sure how AI actually worked.
I recreated your first Tarot GPT results with Microsoft Bing. This is what it came up with:
Sure! I’ll do my best to provide you with a comprehensive reading of your three tarot cards using the past-present-future spread.
The 2 of Swords card represents a decision that needs to be made, but you are feeling stuck and unable to make a choice. You may be feeling like you are in a stalemate or that you are at a crossroads in your life. This card suggests that you need to take some time to reflect on your situation and weigh your options before making a decision.
The Daughter of Cups card represents creativity, intuition, and sensitivity. You may be feeling inspired and ready to start new projects or ideas. This card suggests that you should embrace your creativity and intuition and let them guide you towards your goals.
The 6 of Pentacles card represents generosity, charity, and sharing wealth. This card suggests that you may be experiencing financial success or abundance in your life. You may be feeling generous and wanting to share your wealth with others.
When we look at these three cards together, we can see that they suggest that you are at a crossroads in your life where you need to make a decision (2 of Swords). However, you should trust your intuition and creativity (Daughter of Cups) when making this decision. The 6 of Pentacles suggests that if you make the right decision, you will experience financial success and abundance.
I hope this reading helps! Let me know if there’s anything else I can do for you.
Great article! Most articles I read about using ChatGPT come from copywriters and business writers. I'm happy to see the fiction and literary perspective.
Thanks so much for this positive article. It’s really got useful tips to make AI work for us writers. Looking forward to trying it out.
Bit late this to chat, but to the point several people make about the issue about AI being about the use people make of it in the area of creativity, I'd like to submit an example. Realising as a unknown author, I was unlikely to be interviewed about my book, I got ChatGBT to do it. The result amazed me and might inspire others to think laterally about how to use the tool: https://open.substack.com/pub/invasionlovebirds/p/a-famous-influencer-interviews-the?r=cfm0p&utm_campaign=post&utm_medium=web