UW News

February 25, 2025

Q&A: How AI is changing the film industry

UW News

People picket on a sidewalk.

UW doctoral student Brett Halperin interviewed picketing film workers about AI during the 2023 strikes.Antares Vargas/iStock

In 2023, a good portion of Hollywood went on strike — in part over concerns about artificial intelligence in filmmaking. Now the use of AI has roiled this year’s Academy Awards: Several of the best picture nominees used AI in production. “The Brutalist” showed AI generated architecture blueprints in a scene and its editor used a program called Respeecher to hone actors’ Hungarian pronunciations. “Emelia Peréz” used Respeecher to adjust an actor’s singing voice.

Brett Halperin, a University of Washington doctoral student in human centered design and engineering, interviewed picketing film workers about AI during the 2023 strikes. Their concerns ranged from AI’s effects on wages and jobs to the inauthenticity of the resulting art.

Halperin published the findings Feb. 6 in ACM Transactions on Computer-Human Interaction.

UW News spoke with Halperin about how film workers are thinking about AI and the history of technology in filmmaking.

The striking film workers you spoke with raised various concerns about the use of AI in filmmaking. Were you surprised to see some consternation around the Oscars this year?

Brett Halperin: We have seen backlash to AI from workers and the general public manifest in multiple ways over the past few years — from striking to protesting screenings. Many filmmakers have valid concerns about how studio use of AI can undermine their craft and labor. Meanwhile, many writers and artists object to how their materials are scraped and co-opted as training data for machine learning models without their consent or compensation. This makes AI particularly thorny and controversial. But it’s also important to situate this backlash in the broader historical context.

Throughout history, the “death of cinema” trope has resurfaced with each major technological shift. For example, the use of synchronized sound systems starting in 1926 rendered many silent-era acting techniques, production methods and even professions obsolete. While this caused massive disruption, it ultimately created new professions, such as sound specialists, and transformed rather than eradicated cinema. The rise of color, television, digital media and so on follow similar trajectories. AI presents another iteration of this trope that continues to reflect the shifting cultural and industrial anxieties about technological agency. Part of what makes cinema unique relative to other art forms is that it has always depended on complex, evolving technologies. This change is unsettling, but also an opportunity for all of us, including the Academy, to reevaluate what makes film meaningful.

The Academy is reportedly considering making AI disclosure mandatory for the 2026 Oscars. Do you see value in this?

BH: Generally, I think as much transparency as possible is a good thing. But as AI further integrates into production processes and workflows, excessive mandates could become unreasonably cumbersome and difficult to track. So I would first start by asking: What do we mean by AI? Computer-generated imagery and its associated algorithms have been in the Hollywood studio system since the 1970s. At what point did CGI and other algorithmic tools become rebranded as AI?

In my view, regulation should focus on where AI use has the potential to undermine workers and manipulate viewers. For example, AI actors and de-aging techniques might further intensify body image issues among the public, as well as take work away from actual actors. Disclosure would help the Academy and spectators understand what they are seeing to not only assess the ethics, but also better judge and criticize films in general.

The uses of AI in “The Brutalist” and “Emelia Peréz” are relatively minor. What were workers’ feelings about AI tools as instruments to assist their work, rather than replace it?

BH: The workers did not oppose AI altogether. They seemed to recognize that technological change is an ongoing part of cinema and expressed degrees of openness to the creative possibilities. They acknowledged that there are potentially useful applications insofar as the decision-making power and control over AI lies with them rather than studio executives forcing its integration.

That said, the workers seemed to find current AI-assisted capabilities to be rather unimaginative and unequipped to augment (or replace) their work. For example, a writer who tried to use it to assist him described the written outputs as “hacky” and “generic.” Many of the workers made compelling cases for why AI cannot take over the tasks that truly define filmmaking, such as fostering authentic human connection on and off screen and telling stories that matter to people. 

What were your major takeaways from talking with the film workers? Have those changed at all as the technology has evolved in the last year and a half? 

BH: Despite being around for decades now in various forms, so-called AI today is exhibiting a “novelty effect,” which is currently exploitable, but bound to fade. As AI further integrates and becomes more deeply embedded into cinema like prior technologies, I suspect that the anxiety around it will simmer down.

Rather than fuel the hype cycle, we should remain patient and vigilant in working toward ethical implementations and protections, because AI can incur harms today that require protections for workers and viewers. While Hollywood unions have won protections through collective bargaining agreements, they will need to be continuously updated as the technology  develops, as well as extended to non-unionized workers and workers in other media industries through state and federal policies. I would especially like to see policies that establish informed consent and compensation for artists whose materials are used as AI training data.

What should the public know and consider about AI in filmmaking?

BH: It’s ultimately up to those of us watching movies to decide what we like and don’t like about AI in cinema. We have the power of our attention and wallets to decide what films we want to support. At the end of the day, the Hollywood studio system will invest in what is profitable and divest from what is not. We should listen to the workers for guidance and watch films that align with our values. Despite the current anxiety around AI and the lure of its spectacle today, the public should remember what makes a film truly valuable: the human hearts and souls behind it.

Daniela K. Rosner, a UW professor of human centered design and engineering, is the co-author on the journal article. This research was funded by the Labor Research and Action Network and the National Science Foundation.

For more information, contact Halperin at bhalp@uw.edu.

Tag(s):