The topic of artificial intelligence and its place in the artistic world is a debate that will not go away, and for good reason at that. We should not be replacing human energy and efforts with whatever AI generates. It’s even more dubious when we consider what AI learned (effectively stole) to create from. That debate can be further torn open with the latest quagmire: an audiobook with Michael Caine and his AI-generated voice.
Screen Media
Michael Caine & The Odyssey
As you may know, The Odyssey, directed by Christopher Nolan, is coming to theaters on July 17th. Riding on this hype train is ElevenLabs and their audiobook production of The Odyssey, by Homer. With Michael Caine being a frequent collaborator with Nolan in his films, his inclusion as a narrator here is not a coincidence. Not to mention, it just so happens ElevenLabs signed a deal with Caine to license his voice and likeness for AI purposes.
I think this puts part of the project on a different level of ethical quandary as opposed to generative AI use. Now, keep in mind I just want to focus on the use of Caine’s voice. It should be noted that, reportedly, the entire thing is being AI-generated. That’s another matter entirely that leaves me a bit sick to my stomach. The presence of AI Michael Caine is a different point of contention. Here, AI isn’t being used as a shortcut to avoid hiring an artist. Caine licensed his voice to be used in this capacity. It’s also distinctly possible that if he had been asked to do the project, he wouldn’t have been able to. Given he’s 93 years old and retired from acting, he may not feel up for this kind of job.
The Right to Make Creative Choices
With that in mind, let’s look at it from the creative side. If you want Michael Caine to voice your project and you can’t get him, what options do you have? You could hire a soundalike or someone completely different. Or, if he’s licensed his voice to be used via AI, you could do that too. I do feel that if everyone is on the same page and agrees, this should be up to the producer or director’s discretion. They should note, though, in the marketing and credits that it’s AI and not the actual actor.
That final disclosure is vital for the audience. Ultimately, a creator should be free to use the tools that are legally at their disposal. If they have a creative vision and AI Michael Caine is part of that, let them have it. Disclosing it then lets the audience decide if they want to partake in this or not. I sure as hell don’t want to listen to fake Michael Caine narrate The Odyssey. Hell, I wouldn’t want to listen to an impressionist narrate it as Michael Caine; it would take me out of the experience. The creators have to be cognizant of that, too. They should know that it’s a risk using an AI copy of Caine’s voice and assume those risks.
It’s Not the First Use of AI Voiceover
Let’s compare this to, say, Disney’s Obi-Wan Kenobi series. The voice of Darth Vader was provided by a program called Respeecher, which James Earl Jones authorized. Would I rather hear a synthesized James Earl Jones over a replacement actor? Personally, yes, as long as it’s being disclosed. That way, I know going in what to expect. And if that AI generation is terrible, then I know what and who to blame, including myself for supporting it with my money.
I highly doubt that a projected 13-hour AI-created version of The Odyssey is going to be worthwhile. I firmly believe that even the best AI will never fully capture the gravitas and pathos of an actual actor. All of us are aging, and all of our talents are fleeting. Eventually, we all succumb to the ravages of time. We can only preserve so much of our talent in ways like recordings and programs. AI can only replicate what it is taught and will not be able to think for itself in that regard. Those are inherent risks with AI performances.
Not Every Use of AI is the Same
Every use of AI needs to be taken on its own merits. When Roger Ebert lost the ability to speak due to cancer, technology was being developed to simulate his voice. Ebert eventually passed away before we had the AI technology of today. He could’ve used that to give voice to his reviews if he wanted to. Country singer Randy Travis used AI to help record a new song in 2025. Travis hasn’t been able to fully speak or sing following a stroke he suffered about a decade ago. Should he be criticized for using technology that allows him to give voice to his songs again?
We should all know that morality is not always cut and dry, black and white. What should be pretty straightforward, though, is disclosing how you’re using AI and to what extent. Consumers need to be able to make informed decisions. It’s one of the reasons why films have disclosures about the treatment of animals on set. Just because you can do something creatively doesn’t mean the audience has to accept it. I don’t have to accept AI Michael Caine, but if you want to, then that’s your decision to make.
