Capturing here the full text of an email interview I did with someone, as I know they are unlikely to use the full text in its entirety, and there are some good things here worth keeping a hold of as the work progresses further.
INTERVIEWER: Tell me about a time when AI made your work better. What about worse?
ME: I’ve used generative AI tools to produce 110 short pulp sci fi books in a year, and they have received global news coverage. These books use AI tools like Claude by Anthropic, ChatGPT and Midjourney to explore world building in a complicated storyverse I call the AI Lore books.
In addition to all the brand new narratives I’ve put using them, exploring AI text and image generation tools has enabled me to go back through something like fifteen years of old partly finished writing and world-building, and integrate all of that into new finished works. These have in turn lead to new books connected to those ones in a networked narrative, where readers tumble down their own rabbit holes as they explore the many interlocking worlds I’ve created with generative AI.
It’s fulfilling to be able to bring all those abandoned creative sparks and impulses and seeds and give them a new chance at life within the context of my AI Lore books.
How many hours did you spend learning to use AI?
Apart from professional and personal responsibilities, I would say I’ve dedicated possibly every minute of free time I’ve had over the past year to figuring out how to use AI tools to tell the kinds of stories I want to tell. Actual hours would have to be high hundreds at the least, though possibly over a thousand. Impossible for me to accurately estimate.
Do you disclose the entirety of your creative process to your boss/team? What parts do you keep hidden?
I don’t have a boss for this creative work, but I do have readers, to whom I feel a certain amount of responsibility in terms of providing a quality product.
My product is interesting short illustrated AI “mini-novels” as one media outlet described them. So I disclose to my readers in a variety of ways on my site and my store that I am an AI publisher, and the works contain AI-generated content. The copyright pages of the books all include notices themselves that aspects of the content may have been AI-generated, with human review and editing.
Some of the books contain more or less AI-generated writing (all the art is 100% AI generated), and there’s no simple means currently to easily indicate in a passage of text to your readers which parts were human-generated, and which were AI-generated. It might be interesting as a sort of “x-ray” mode of a text for e-readers, but I haven’t seen anyone solve the technical problems here yet. And I’m not sure so far how much people reading my books care which words were written or edited by me, and which came out of what LLM text generator tool.
So I leave those questions to future technologies, and focus on simply presenting an interesting and fun creative exploration of the tools, their capabilities, risks and limitations.
What’s one piece of advice you would give to someone in your industry when it comes to working with AI?
I’m aware people have reservations about participating with these technologies for valid reasons, but for me the only way to really understand them and their impacts is to actually use them “in the wild.” This way, I can gain realistic experience and perspective on the problems and possibilities that are offered by the tech in the here and now. If you wait around until technologies are perfect before trying them, you’re going to be waiting a very long time. In my opinion, it’s better to actively participate, have a seat at the table, and help steer the technologies in directions that serve everyone better.
When you generate something using AI, what are you doing to make that work yours?
I’m not concerned with issues of authorship or my identity as “me” when I put together these books. I have a premise usually I’m exploring, or a topic from another work I want to expand on. So I simply follow what feels true to the creative thing that is trying to express itself in the moment I am creating that book. So generally, I produce too much content at first, and then as I see what is generated, and analyze my own reaction to it, and whether it feels “true” to that world, or not, I further narrow down what will be included.
So it is a process of first ideation usually, then content generation, then reduction/selection/curation. Followed by arrangement into a linear form in the ebook, and editing, cross-linking to other volumes, and any other improvements. In this process, even if I don’t anywhere identify myself as the “author,” I believe still my choices as a person shine through and make the work plainly “mine” as opposed to someone else with an entirely different life context using the same premise, but ending up in a totally different place than where I took it.
For me the creative process of using AI is more like being a producer, or a creative director, or even a film director. You make decisions about the what, and the how, but you leave many of the fine detail work to the skill of others, and the capacities of external tools. It’s your vision that’s carried out through all the choices you make along the way, and all the creativity and imagination you can bring to bear in the moment. And capturing that spirit of the thing that moved you to create this work. It’s so intensely personal to me that I don’t feel the need as an author/artist/creator to “own” it. I merely feel the responsibility to represent it as accurately as it appears to me, and try in my own way to communicate it to others.
I think the job of the artist here becomes assembling the nexus of components whereby the work can take place in the experience of oneself and others. Both artist and audience bring their own contributions and recipes to the final mix. In a way, the artist becomes just an advanced member of the audience, having gotten there first, and left some signs and markers along the way.