I ran an experiment. The results were far more interesting than I could have imagined. I’ve been working with Artificial Intelligence for a number of years, among the early adopters of ChatGPT, and then others as they came around. I used them for all sorts of things, from dinner recipes to inquiries about quantum consciousness. I asked questions, exquisite and complex prompts to work out pieces of .Net code, including JavaScript, Python, design patterns, architecture theory, and even user experience.
At times, the results were intricate in ways I didn’t anticipate; other times, it required a great deal of refactoring to make it compile.
As an author, I used it for research, character development, and plot tactics. I wondered about reader expectations, what techniques were working in the markets, what was overdone, and conversely, underappreciated.
I look at it this way: Artificial Intelligence was present in my workspace. It did not hold the pen.
In the decade since I began authoring in earnest, I’ve written and self-published eight novels. A short story found its way into an anthology published by the Utah chapter of the Horror Writers Association. I created a recording studio in a corner of a spare bedroom where I practiced the art of voice acting. Hundreds of hours were spent reading manuscripts, editing the wave forms, and publishing the audio versions of a number of novels.
In other words, I was a one-person creation studio. I penned the words, I gave them a voice, I edited, perfected, and produced creative artifacts. I marketed them. I stood in the wind, rain, cold, and heat to peddle these wares to fiction and fantasy enthusiasts with all the polish of a professional.
I am aware of how the advancements of technology can affect every aspect of the creative industries. Human innovations exceed earlier predictions in unprecedented ways, and I continue to marvel at the ingenuity, the level of sophistication, and the incredible speed at which progress advances.
Mind-boggling as it appears, platforms can receive a prompt, research the topic, plan, compose, and produce an audio-visual report of their findings. All that labor, hundreds of hours of human labor, reduced to a handful of minutes of automation. I am deeply concerned for the financial future of fellow creatives.
For that matter, there isn’t an occupation that won’t be affected by artificial advancements. These technologies are here, and whether or not we’re prepared for the future, they will not disappear simply because we fear for the future.
So, I conducted an experiment. I downloaded an open-source Text-to-Speech AI model that I discovered, ironically enough, through one of my chat investigations.
It required a small amount of programming to take an entire full-length novel, break it down into chapters, paragraphs, and sentence fragments short enough to work with the model. Then, in a loop, the program I built started at the beginning, fed a bit of text along with a six-second sample of my voice, compiled a data stream, and produced every chapter of my novel as an audio wave file.
Because I am a nerd, I kept track of everything I did to produce an audiobook from the completion of the print version to the distribution platform. All told, roughly speaking, for every minute of listening to one of my stories, I spent just over twenty minutes recording, editing, and producing the audio.
What would take me months of labor, fitting in recording sessions between work, family, and other obligations, I accomplished with a couple of hours spent writing software. In a single night, while I slept, my voice was used by Artificial Intelligence to produce an entire full-length audiobook.
In all honesty, I am unlikely to distribute these files. While undecided, I find myself leaning in this direction because it feels like I’ve lost the artistry, that human-only factor that makes listening an experience. What AI gave me was narration.
Don’t get me wrong, the model included intakes of air at appropriate spots. It inserted nuanced pitch rises and falls that made the narration less artificial. It breathed in all the right places. It felt nothing. The output lacked what great narrators give to their performances. It’s why, when I produce an audiobook, I include the phrase, “written and performed by” along with the title.
In the end, my experiment produced fascinating results. It taught me a ton about what Artificial Intelligence can do. Perhaps, somewhere down the road, I’ll have to repeat and reanalyze. But for now, I think I’ll stick to the performances, as daunting as that sounds. Because narration might fill the silence, but performance earns it.
Stories That Bring Technology To Life 