The first time I heard voices I thought I was losing my mind. AI-generated voices, that is. Because I couldn't believe how "real" they sounded. Listen to the intro to this episode and tell me what YOU think. Then find the little samples of producer Jim Mullen's voice later in this episode!
I'm still wrapping my head around all that is possible -- and all that has changed in the world of content creation and media production. That's exactly why I knew I had to bring together two of the most credentialed producers I know to talk about what's really happening behind the scenes with AI in our industry.
The Real vs. The AI - A Producer's Perspective
Jim Mullen (Mullen Media) and Yossi Kimberg are the real deal – we're talking Seinfeld, Food Network, Super Bowl halftime shows, and decades of work on Lorne Michaels shows. These guys aren't just talking theory; they're actually using these tools in production right now.
AI-Generated Yossi Kimberg
AI-Generated Jim Mullen
What Really Struck Me
Yossi's journey into AI started by making masterful still images with Leonardo, as in Leonardo.AI. As he put it:
"The idea that text generates an image was beyond imagination. I was just amazed by it."
But then he progressed to motion, face-swapping himself into different editing scenarios (find those in this episode, too!), and even bringing children's artwork to life. But learn what happened next that was "a bridge too far" for one particular client.
Now, while Yossi's moved on to scripted motion videos all AI-generated, Jim's been experimenting with ElevenLabs for voiceovers, though he hit a hilarious snag on a project about a lack of libraries:
"I couldn't get the bot to say 'desert'. It kept saying, New York City has book desserts."
Not so funny to me, because once it learns contextual pronunciations, my gigs doing voiceover will dry up like a desert. (Note: I found out just today while producing this episode that I literally lost a job to an AI-generated voice. True story.)
The Uncomfortable Truth
Both producers agree we're not just approaching a tipping point – we're already there. Yossi believes that within 2-3 years, even the very role of managing AI might become obsolete. The technology is learning to be funny, dramatic, and emotional without human instruction.
Jim pointed out something that made my stomach drop: if you can create an entire commercial with AI actors and voices (which is already happening), you eliminate hair and makeup artists, grips, gaffers, location scouts – entire production crews become unnecessary.
As such, we talked about the fear factor. You know how I feel about voiceovers. Though AI helped me immensely in honing these show notes. But it's not just job displacement we're facing; it's questioning our creative thumbprint and what makes us uniquely human in this process. Plus...
Copyright and Ownership Minefield
We touched on some fascinating legal territory that's still being sorted out. Yossi mentioned reading legal briefs about using copyrighted images in AI training, and the emerging trend of celebrity estates licensing deceased stars' likenesses for AI projects – Steve McQueen's estate is apparently leasing his character for AI-generated content. The whole landscape of intellectual property, name-image-likeness rights, and even the ability to deny you said something ("I didn't say that, it was AI") creates a complicated web that we're all navigating in real time.
(I recommend taking a listen to the brilliant Robert Tercek who allays some of those legal fears in this Insider Interviews episode from last year: https://bit.ly/Insider21-Shapiro-Tercek .)
The Human Pushback
But here's what gives me hope: Yossi shared how his client loved his AI-enhanced children's artwork until clients pushed back, saying it wasn't authentic to what the kids actually created. They pulled back some of the AI elements. It reminded us of the vinyl comeback – sometimes we crave imperfection and...