#49 - Artificial Thoughts

Last week I had the opportunity to sit down with a group of fellow authors and just talk about all the curious aspects of Artificial Intelligence. We were tackling the question of how to think about it, plan for its effects on our businesses, and most of all the ethics of AI. Utterly fascinating, so I thought I'd share a bit of what we discussed.

AI really breaks down into 3 distinctly different things: AI Tools, Generative AI, and AGI (Artificial General Intelligence). 

AI TOOLS: If you use a smartphone, do a search on a website, or spellcheck your email, you're probably using an AI Tool.  If you are tracking a package from Amazon and it tells you when it will be delivered, and maybe in how many more stops, you're using AI. Some are more powerful, some less so.

GENERATIVE AI: This is the one that is in every author's mind, an AI that can generate text (as in: good, readable, enjoyable story). Frankly, it ain't here yet. Oh, there are folks using AI to write their books and publishing them, but it just isn't prime time yet, no matter what some authors are saying. Is it coming? Eventually...maybe. But most folks are still convinced, myself included, that there is a creativity and imagination that simply won't be able to do anytime soon. Also, learning how to tell and shape story takes an immense amount of practice. Skills that generative AI simply can't come close to yet without an already skilled storyteller driving the prompts...and maybe not even then (yet).

AGI: This is the monster of the science fiction movies, and we aren't there yet. Will we be? Sure. Soon? Maybe. Will it kill us all? Unlikely. Think collaborative partner that's way faster at thinking than we are and you're probably much closer to the truth.

BACK TO AI TOOLS: This is where author discussions are really happening. There are many different levels at which we can use AI tools. Here are a few. I've put them in kind of an escalating order of ethical complexity

• Complex research questions: AI will dig around and find all sorts of interesting information much faster than a series of searches. This is very handy.
• Story bibles, etc.: NotebookLM allows us to load a book or series into the AI, which we can then ask questions of: "What color are Kate's eyes?" "When was the first time these two characters kissed?" Or even more useful, "Summarize each of these books, listing the main plot points and character traits." And maybe, "What are the typical beats and rhythms of my thriller series?" Something that would be very time consuming for someone to catalog, AI is starting to be able to do. One of the questions I have though is, "If I know the patterns of my own writing, will that interfere with my creating my next book?" This is nearing where I draw the line.
• Using simple Generative AI to create blurbs and other marketing materials. It still needs a good human touch, but could certainly save time.
• AI audiobooks. Human-narrated audiobooks are very expensive and/or time-consuming to produce. AI Audio is really taking off. For the most part it's a little flat and dull, but it tells the story. There are massive innovators in this area and I'm certainly experimenting with https://Spoken.Press as one of them. This is dancing along an ethical edge. Most AI audio, the voice was created by teaching the AI on other people's voices. Spoken uses AI voices that were licensed by the voice creator who then receives income every time their voice is used. The former bothers me, but the latter? I'm glad to save time and support a voice artist.
• AI imagery: This gets tricky. AI can make beautiful cover art, even stunning with the right prompting by the user. However, these imagery engines were trained on something...the art of thousands of other artists, without credit or renumeration. There are already numerous lawsuits on-going about this.
• Problem-solving: Authors get stuck when they're writing, a lot. Some descend into full-on writer's block. Many authors will use AI as a brainstorming partner. "Here's what's happening in my present work-in-progress. I'm stuck on what this character is likely to do next. Please give me five ideas of where I could take this story." I'm not a big fan of this. I think that my own writing and brainstorming with my wife (who's really good at it) creates a more unique tale. Still, I think the ethical questions here are fairly minimal.
• Outlining: "Look at my six prior books in the series and all of the bestsellers out there. Write me an outline for the next book in the series that will mirror the beats of a bestseller." Sorry, nope, not for me. Yes, I read other authors and learn from them. But this shuts out my own creativity and gives it to an AI that was trained on other books. Many of my own books were illegally used to help train the AIs that would try to answer that question. Tens of thousands of books have been identified as part of this. Nope.
• Writing the book: Excuse me while I laugh. Sure, one might do well here and there, but innovating, creating new story, engaging at all but the most mundane level? Not gonna happen. ...at least not yet!

This whole post could have been generated using AI. Would it have sounded like me? Probably, if I prompted the AI to base it on my other work. Would it have included my thoughts and opinions on AI usage? Not so much. It could have created a discussion with examples, but it wouldn't have been me or mine. (No AI was used in the creation of this post, not even for basic research. [grin])