I have continued to wrestle with the implications of AI in my industry (tabletop RPG design) given that AI both threatens art and writing. The technology may not quite be there yet, but with every iteration and update the software gets better at hands and better at checking its own work.
I don’t want artists to be made redundant.
I don’t want my work as a writer and designer to become redundant.
I want to continue to employ artists and writers as well.
Things are further complicated by the various stances of companies such as Paizo, DrivethruRPG and others. There’s a great deal of hostility on both sides of the argument, between the Cult of the Machine on the one hand and the neo-Luddites on the other.
Ultimately, in a business with margins as narrow as the TTRPG industry, AI is always going to be a temptation. I have zero reservations about using AI to produce YouTube thumbnails and other, similar ephemera like live stream backgrounds or blog art, but when it comes to illustrating books things become a bit more complicated.
Other small press companies are using AI to great effect. Before The Red Room were shitcanned by DTRPG they used a mix of art, often a lot of it AI, which enabled them to rapidly produce material and to earn a great deal of money compared to other small companies (according to the public details). In no small part they were able (and are able to) produce content at such a rapid pace because they don’t have to necessarily wait for art. Other small publishers have followed suit.
If I am to compete (in the general rather than the specific sense) I have to follow that path, to at least some extent. If I can turn out product more swiftly and at reduced costs, I can make more money, and more profit, and invest that profit into the products I have a lot of faith in.
Yet I don’t want to go all the way that way either.
So I’ve decided to double-down on the ‘Radical Centrism’ on this topic. I’ll use AI where it makes sense to do so – filler pieces of art, rounding out what I have to add richness and quality to work. I’ll also use AI on projects where I think there’s less potential popularity, less potential profit and where I might not have otherwise been able to make the project viable. These projects, tester projects, will be tagged or in sections called ‘Rapid Prototype’/ If those projects become sufficiently popular, I’ll then return to them and give them the full, proper treatment in a proper edition.
Main works will still be predominantly human-driven.
We’re all struggling with this and figuring out where we want to be with it all, how to integrate it without undermining human creativity. Hopefully this is a step on that path.