I started playing with Voice.AI the other day. As of this writing, the product is in Beta. It takes a recording or input from your microphone and transforms the speaker's voice into another speaker of your choice. Want to sound like H. Jon Benjamin or Morgan Freeman? With this program, it is as easy as record and upload.
Currently, Audible does not allow AI to record audiobooks. It explicitly states that "ACX doesn't accept text-to-speech or other automated recordings." Does that apply to human recordings transformed by AI? Not sure, but I'm assuming as much when producing my audiobooks. It's the safer assumption, not only for not violating ACX's submission requirements but for other reasons I'll get to in a bit.
Why AI voices would be awesome!
But what if we could? When I created the robot for Intergalactic Space Force, I pictured his voice as the robot from Lost In Space. Even some of his design elements come from both the classic version and the redesign for the 1998 movie, including the transparent torus-shaped head and arm cannons, but I had his voice in my head when writing the book. The first scene where the robot meets our heroes, only to turn to them and say, "You are all my bitches now" - in my head, I hear it in Dick Tufeld's deep, staccato voice, and I chuckle every time. When I record the robot, I try to capture that same robot voice, but I'm no Dick Tufeld. What if I could, though? What if I could upload my recording and transform that dialogue to sound exactly like I imagined? In the sequel, we meet a clone of Elvis Presley (just go with it). What if, instead of my poor impersonation of Elvis, we transformed my dialogue to sound exactly like Elvis? How cool would that be?
The issues with AI-generated voices
And how badly would I get sued? That's the other shoe. Where do we draw that line? What are the legal ramifications of stealing another person's voice? For that matter, how is it different that me impersonating that person? Isn't this just a better, computer-aided impersonation? We are entering a new era of AI-generated content, not only for voices but for cover art and book content, and the laws have yet to catch up.
For now, it's a fun toy to play with. I'll hold off using it in audiobook productions . . . for now.
For cover art? That's a different conversation . . .