Sounding an $800-Million Alarm on AI

Tyler Perry in a tux posing in front of a Time Magazine red backdrop

Movie mogul Tyler Perry halted an $800-million expansion to his Atlanta movie studio lands last week. The former army base is already one of the biggest film studios in America. Plans called for a dozen new soundstages and more to be built on the 330-acre lot. But after testing a new text-to-video AI program, Perry pulled the plug.

So, what does this have to do with regular people? A lot, I think. Especially for regular creative people who make their living in content creation.

When voice-over work began a noticeable industry-wide slowdown last year, many of us in the industry talked about the reasons why. AI- Artificial Intelligence – came up over and over again. Creative project leaders and company executives wanted to wait and see whether they’d be able to do things cheaply with AI.

Undoubtedly, if a one-time purchase for an AI program can yield spectacular video, why pay people to make video? Once AI voice-overs are able to replicate human emotions and cadence, there won’t be a need to hire a human for some projects. People want lunch breaks and decent wages and health benefits. AI needs none of that.

AI Isn’t There Yet

Early this year, I wrote about my judgment error in accepting an audiobook project that turned out to be written by AI. Last week, a radio station was roasted on Facebook for obviously using AI to write blurbs about its on-air personalities. AI “writes” in generic and bland terms. And we can all still tell when an AI voice is replacing a human. But it will get better.

There’s a perception that voice-over is expensive. I’ve said this before – it’s partly because the end product sounds effortless. What went into creating that smooth, silky, perfect piece of audio is largly hidden. Last year, a regular client told me that voice-over was “so pricey” that it was taking a big hit out of her profit margins. I wanted to keep her and she wanted us to keep working together. But she had a valid concern, especially about the rate for “pick ups” – when the client makes small script changes and they need new audio. So, we renegotiated this cost. I took her on faith and she was relieved that I was willing to budge a little. We have done many more projects together.

Last week, I had a discussion with a potential client about narrating their audiobook. They were horrified that I wouldn’t take the project for $125 Per Finished Hour (PFH). My rate is higher. To the uninitiated it might seem like I’m looking for an unreasonably high hourly rate. Finished is the key word. Each hour of finished audio for an audiobook takes about six hours to create. I’m certainly not thumbing my nose at $21 per hour, but with my experience, I charge more. And if someone doesn’t want to pay it, I’m happy to let them walk away. I’ve come too far to do work that makes me feel undervalued.

BTW script changes are different than pick-ups to correct any errors I might make – those are always free.

Back to Tyler Perry

The poster for Perry's Boo 2 a Madea Halloween movie with Perry dressed as his old lady character, Madea.

Billionaire Perry was so blown away by a demonstration of Sora, a new text-to-video AI program from the makers of Chat GPT, that he put his massive studio expansion on hold. The program, he says, can eliminate the need for travel to locations, for building sets, and he has even bypassed the makeup chair because AI can make him look older for a role.

But as an actor, he’s also sounding the alarm about what AI will do to the film industry. Last year’s massive actor’s strike finally resolved after an agreement about how AI will be incorporated alongside live actors, not instead of them. For example, if an AI-generated version of an actor is used in a scene, the live person will get paid. AI will be a tool, not a replacement, for writers and others.

But for how long? And what about set designers? Location scouts? Make-up artists? And how about the rest of us? Someone also needs to educate us on how to spot the differences between “real” video and the fakes.

Perry says Congress needs to incorporate these agreements into law, not just in various union deals.

I think we’re pretty much on our own in voice-over. We have an advocacy group, but it’s in its infancy. Joining a union in Canada is prohibitive. I used to belong to one and it prevented me from working because most VO work is non-union. Unions view non-union work as undercutting rates, but I view it as living in reality. My rate is my rate but I can also be flexible, when I might lose a long-term client. Or steadfast when it’s right. You can’t negotiate with AI. But the advantage for users is, you’ll never have to.

1 thought on “Sounding an $800-Million Alarm on AI”

  1. Hi Lisa. I am enjoying your posts about the inside dope of VO work. We both work in industries where we provide a talent service that is often viewed as a common commodity by clients who don’t understand its value on the audience. We provide a specialized service – live, experienced talent.

    I have clients who seek me out and are able to pay what we think is my value. Other clients value my work yet cannot pay but offer repeat bookings. At the bottom are clients who don’t care who/what is in front of their audience so long as it fills the time needed. These clients offer the least. I actively seek out the former two, and accept the latter only if I can see a business advantage by doing so.

    Also, a few years back I stopped worrying about the short term dollar vs. long term gain. My ‘time’ is worth nothing to the client, only to me. Clients only care about the end result and its value to their audience. If they don’t want to pay for my experience and my value, I view the client’s reasoning as not caring about their audience. There is an agent who regularly hooks clients for top dollar and then tries to hire out for less than 50% of what the client is paying. No one will work for him because he doesn’t value the performers he is trying to hire.

    A similar issue in my business is that you can buy a career. You can go to a magic store, buy a trick, learn it, and get paid for showing it that night. This is what I would consider the magic version of AI.

    Will the audience know the difference? Some will. Will they care? I don’t know but what I do know is that I don’t.

    There will always be a market for quality ‘real live’ talent. AI generated entertainment is the fast-food segment of the industry. There will always be a market for it however there will always be a market for a really good steak.

Leave a Comment

Your email address will not be published. Required fields are marked *