By 2023, artificial intelligence had seeped into enough corners of a nervous film industry — buoyed by Barbenheimer, but fully aware of an imminent 2024 shortage of new titles — to become a seriously effective tool of labor unrest. Last year's Screen Actors Guild contract, achieved after a lengthy, costly staring contest with industry producers and streamer honchos, added some guardrails designed to protect actors' collective livelihood, noting "the importance of human performance in motion pictures and (AI's) potential impact on employment."
Tye Sheridan knows about that impact. He's an actor, having made a formidable screen debut in the 2011 Terrence Malick film "The Tree of Life." He's best known for Steven Spielberg's "Ready Player One" and as Cyclops in the "X-Men" movies.
Sheridan is also really into AI. He co-founded Wonder Dynamics in 2017 with his partner, visual effects supervisor and filmmaker Nikola Todorovic. They now oversee 70 employees in the U.S. and in Todorvoric's native Serbia.
What is Wonder Dynamics, besides a name promising both wonder and dynamism?
Its founders say it's an affordable, easy-to-use shortcut for filmmakers with projects calling for computer-generated characters. The AI platform (monthly subscriptions start at $20) offers the user a variety of characters. A robot. An alien. A bearded professor, with the slump-shouldered, underpaid air of the average adjunct.
Let's say your screenplay calls for a shot of your alien running out of a building, stopping, looking both ways with a worried expression, and then running off again. In a real location, you film your real actor, running. You then take that raw footage and, with the Wonder Studio software, you turn your human into an alien, without any pricey motion-capture suits or lengthy post-production effects phase.
Todorovic and Sheridan have many fans and customers, including the Russo Brothers ("Avengers: Endgame," "The Gray Man"). Joe Russo is on the Wonder Dynamics advisory board. The Russos hired Wonder to work on their next project, "The Electric State," due in late 2024 or early 2025.
They also have their fair share of skeptics. One LinkedIn commenter said this of the Wonder Studio AI: "Someday this will be reverse. They will film a robot and use AI to bring back a legend from the grave." Another said: "Say hello to job loss as well."
https://vimeo.com/805754885
In the wake of a Chicago Humanities Festival event featuring them, I spoke with Todorovic and Sheridan to get my head around the implications of what they're selling. Our conversation has been edited for clarity and length.
Q: Consider me AI-agnostic at best, Tye. What's your sales pitch?
Sheridan: Our main goal, really, is to create opportunities for artists. To allow stories that were completely unattainable to most filmmakers to be told. We're trying to bridge a gap, reduce (filmmaking) costs, reduce the time it takes, for the industry and the creators. The audience just wants a good story, and there's a lot of good storytellers out there that haven't had the opportunity to tell their stories.
Q: Nikola, at the Chicago Humanities Festival talk, you explained your company's approach to some of the ethical questions. You said that you're not doing generative AI, you're "just extracting (visual) information, observing actors moving and performing." Can you elaborate on that?
Todorovic: We're accelerating artists' work, not generating it ourselves. We don't train our models on existing art. We work with what the artists have made. We don't really want a future where there's no performance, or there's no cinematography as we know it, and it's all getting digitalized.
Q: So you're saying you're on solid ethical ground because you're not leaving a human performer completely out of the process? That you're selling the equivalent of responsibly sourced ingredients?
Todorovic: Yes. I mean, Tye is an actor. We don't want to put him out of work.
Q: The other night you acknowledged the possibility of a film production future where shooting a movie in person, on a set, has become passé, or at least as rare as a black-and-white film. I guess I'm nervous about where this might go without guardrails. So. Assuage my fears.
Todorovic: It is a fear of ours as well. I hope that future doesn't happen. We have to have performers, and performance art, which is a huge part of the magic of storytelling. Even if the future means (computer) generating certain environments, the performance in my mind is still going to be performed by an actor. Otherwise it's hard to generate characters with feelings. That's why we're building our company holistically, and keeping it in the 3D space. That way the result is always going to be only as good as what your cinematographer did with the light, in real space.
The audience will tell us where it all goes in the future. Will the audience like watching something completely synthetic? I don't think so. We love to watch other people. At Wonder Dynamics we don't want to be part of the wrong kind of future.
Sheridan: It really does come down to the audience. They'll dictate the stories we tell, and what the medium becomes. You also have to consider the economics. We've seen them change a lot in the last five, 10 years. Theatrical distribution has completely shifted, and the economics of making certain films has changed. I'm not talking about the "Avatars," but about the films getting pushed out of the industry because not enough people are going to see them.
Q: I have to assume that coming off a long strike, as both an actor and the co-founder of an AI platform company, you've taken some (heat) from some actor colleagues about this sort of technology.
Sheridan: When something has the potential to shake up our industry in a fundamental way, like AI, it's our natural impulse to get defensive. If you're an actor you're afraid that what you value, what you have to offer, won't be valuable in the future. People tend to jump to an extreme reaction and say (AI) is going to replace everyone. We definitely heard that during the strikes.
I also heard a lot of folks who saw the benefits, and there are some people trying to pump the brakes a little (on the anti-AI rhetoric). Saying AI is bad or good, that's too general. It's like saying the internet is either one or the other.
Q: So how do you stay on the right side of the ethical line, when your technology could so easily go in the direction you say you don't want effects-driven filmmaking, or filmmaking of any kind, to go?
Tudorovic: Every time we add a new feature, we have to use our compass to see if it will affect our ethical mission. You're 100% right. This is probably something our investors wouldn't like to hear, but yeah, we are tempted a lot. New research comes out, and you think, "Oh, this would be so cool, this is super flashy, this would be amazing for social media! A lot of the (generative AI companies) just want to build tools for social media, where they have billions of users. Instead of building tools for a few creatives, which is what we're doing. I mean, you're tempted to package something that could be a quick and easy new feature that gets you millions more users. But Tye and I don't want to build something that we, as artists, don't respect.
The fear comes from where we'll all be in three to five years.
Phillips is a Tribune critic.
No comments:
Post a Comment