Column: AI is coming for Hollywood

The looming threat of AI in Hollywood will arrive with all the subtlety of the Kool-Aid Man busting through a wall.

Artificial Intelligence is primed to take over Hollywood with all the subtlety of the Kool-Aid Man busting through a wall.

Proponents of AI are making bold promises. That potentially means human creativity will be replaced by the theft — sorry, data scraping — of pre-existing words, sounds, images and ideas. Jobs will be lost as human actors and crew are eliminated from the process. Clean water will be wasted, with billions of gallons needed to cool data centers. What will be left is anyone’s guess.

Despite these concerns, AI has found (bought?) a berth at film festivals, which purportedly exist to celebrate the art of cinema. At Cannes this year, a producer was hawking AI translations of international films. Actors who make a living dubbing such films? Soon to be obsolete apparently.

Earlier this month, a Korean film festival in the city of Bucheon launched its first competition dedicated to AI filmmaking. And closer to home, the Tribeca Film Festival in New York featured a program of short films made with generative AI.

Among the filmmakers taking part was Nikyatu Jusu, the writer-director of the horror feature “Nanny,” which won the Grand Jury prize at Sundance in 2022. My colleague Michael Phillips called it an “eerie, assured feature film debut,” and it also got a home video release from the prestigious Criterion Collection. All of this is background to say: Jusu is respected and her work is well-regarded. So it came as a shock to many that she was embracing AI.

That’s because she expressed her own concerns a year ago on social media: “Can’t stop thinking about all the various ways AI will be used to replace living, breathing minoritized artists already struggling to tell their own stories. AI will become mock representation — an empty mimicry of Black and brown people’s light commodified and vomited back to us.”

These are valid concerns. But last week she was promoting her Tribeca premiere on Twitter with a still image from the film and the caption: “Crucify me now, get it out of your system.” Some people engaged in good faith. “Not an attempt to crucify: What made the plagiarism tool attractive as a filmmaker?” someone said. “Ask the people who created it and studio execs who are currently implementing it,” she replied. “Odd to ask the creatives at the bottom adapting to inevitable change … this is here whether you kick and scream or not.”

I don’t know that AI-made films are inevitable, but let’s table that for a moment.

“This is so depressing,” someone else said. “I know,” she replied facetiously, “so many Black faces.”

But of course, none of the faces in her film belong to actual Black people. Those images were constructed from real people who did not give their consent, nor did they receive compensation. How is that a win for representation? To quote Jusu’s own words, this is “mock representation” and “empty mimicry.”

Others were less polite in their comments and Jusu eventually deactivated her account. I reached out to see if she would be open to talking but did not get a response. Some questions I would have asked: What was it like making a film this way? Was it creatively fulfilling? What did she learn from using this technology? Is she allowed to talk freely about the experience, or was she asked to sign an NDA prohibiting her from discussing the tech in anything but glowing terms?

People make all kinds of decisions for reasons we aren’t privy to. Sometimes there are financial considerations. Sometimes people simply rethink where they stand: If you can’t beat ‘em, join ‘em. So I would have also asked: What changed her mind about AI?

I hope Jusu is given grace for what might have been a decision made under stressful conditions — funding and jobs have dried up for many in Hollywood at the moment, where everyone is encouraged to “stay alive until ’25” — but I also hope that, going forward, she gets more opportunities to make films the traditional way.

But we should be wary of AI for ethical and philosophical reasons alone. According to the Human Rights Watch, “personal photos of Brazilian children are being used to create powerful artificial intelligence tools without the children’s knowledge or consent.” According to another report, Google image search is “serving users AI-generated images of celebrities in swimsuits and not indicating that the images are AI-generated” and in some cases, those celebrities “are made to look like underage children.”

Even from a craven capitalistic point of view, does it seem odd that studios aren’t more vocal about potential copyright infringement?

There are aesthetic concerns as well.

As of last year, Robert Zemeckis had plans for a movie starring Tom Hanks and Robin Wright that will use “new hyperrealistic technology including AI-generated face replacements and de-aging to allow its stars to tell a story that spans generations.”

Directors used to simply cast other actors to play different versions of the same character — and happily, many still do. That’s one of the more enjoyable aspects of “Interview with the Vampire” on AMC. The show didn’t opt for de-aging software to make 71-year-old Eric Bogosian look younger for scenes set in the past, but hired 36-year-old Luke Brandon Field, who looks a lot like Bogosian and has the talent to emulate his performance. One of my favorite examples is 11-year-old Mayim Bialik playing the younger version of Bette Midler’s character in 1988’s “Beaches.” As a viewer, I don’t want to lose that. I don’t want actors to lose those opportunities, either.

There are other potential uses for AI floating around, one of which was recently touted by Ashton Kutcher: “You’ll just come up with an idea for a movie, then it will write the script, then you’ll input the script into the video generator and it will generate the movie. Instead of watching some movie that somebody else came up with, I can just generate and then watch my own movie.”

I can not imagine anything worse. After a long day, I have to do more work just to watch something entertaining? How bleak! “What if each time you watched a movie, it played out differently?” teased another news report. “The idea is to use AI to mix up scenes and create completely different versions of the same movie each time it is played.”

These are terrible ideas.

“Going all the way back to gathering around the fire, we like to be together to tell stories, it’s important to us,” technologist and media analyst Sydette Harry told me. “Cinema was another physical manifestation of that. Social media is the digital manifestation.”

The word community gets tossed around a lot and may have been flattened in the process. But experiencing entertainment together can be a cultural bonding agent, which has been fractured by the binge-release model on streaming. These proposed uses for AI will fracture the experience even further.

“Another thing about AI,” Harry said, “is that it likes to put a definitive stamp on things: This is ‘the knowledge’ because we’ve scraped all there is to know. Well, we really don’t know all there is to know.” There is an entire history of existence that has not been digitized, which means as far as AI is concerned, it doesn’t exist.

There’s an assumption that if you’re resistant to AI, you don’t want progress, Harry said. “But if you want creative progress, you have to leave space for things to be sloppy” and allow for ideas to spring from imperfections and happy accidents. “Think about the things in cinema that are really amazing. If you were in a real gunfight, the one thing you never want to do is the traditional John Woo limbs-akimbo pose. But it looks amazing on screen and now it’s a signature thing dreamed up by an amazing filmmaker.”

“People’s discomfort with the human learning process — and the ability to play around — should not become the guide rails for our art forms,” she said. “We have to allow for it if you want any kind of culture. Putting it through a computer that has weighted judgments about what’s important? Well, that makes somebody in an executive office feel good and then nobody has to be accountable for what a show or film is saying because the black box of the computer said it.”

Even if AI is a cheaper way to make films initially, I suspect that will change; once studios become reliant on it, prices will go up (not unlike Uber’s pricing trajectory) but it will be harder for studios to go back to the “old” way because fewer people will have the necessary skills anymore.

It’s worth hearing out the concerns of people who don’t have a financial stake in AI’s domination. Privacy expert Meredith Whittaker recently described the AI business model this way: “It costs hundreds of millions of dollars to train these models, so there is deep pressure from companies — that are basically promising God and delivering email prompts — to make some return on investment in this technology.”

Timnit Gebru is a researcher in AI ethics and she is less convinced the technology is a fait accompli: “This ‘it is inevitable’ discourse is designed to disempower. It’s not inevitable. These dudes don’t need to be handed ridiculous amounts of money to realize their dystopian utopia.”

And according to at least one recent headline: “Payoff from AI projects is ‘dismal’, biz leaders complain.” Despite the marketing and propaganda, AI isn’t even the golden goose.

Nina Metz is a Tribune critic.