q&a:-how-ai-is-changing-the-film-industry
Individuals protest on a sidewalk.

UW doctoral candidate Brett Halperin engaged with protesting film workers about AI amid the 2023 strikes.Antares Vargas/iStock

In 2023, a significant segment of Hollywood went on strike — partly due to worries regarding artificial intelligence in filmmaking. The incorporation of AI has stirred controversy at this year’s Academy Awards: Many of the best picture contenders integrated AI during production. “The Brutalist” featured AI-generated architectural designs in a scene, and its editor utilized a software called Respeecher to refine performers’ Hungarian pronunciations. “Emelia Peréz” employed Respeecher to modify a performer’s singing voice.

Brett Halperin, a doctoral candidate at the University of Washington in human-centered design and engineering, interviewed striking film workers concerning AI during the 2023 protests. Their apprehensions spanned from AI’s impact on salaries and employment to the artificiality of the resultant art.

Halperin published his insights on February 6 in ACM Transactions on Computer-Human Interaction.

UW News interviewed Halperin about the perspectives of film workers regarding AI and the technological history within the film industry.

The striking film workers you conversed with highlighted various apprehensions regarding AI’s role in filmmaking. Were you taken aback by any discontent surrounding the Oscars this year?

Brett Halperin: We have observed backlash against AI from both laborers and the public being demonstrated in various forms over recent years — from strikes to protesting film screenings. Numerous filmmakers have legitimate worries about how studio utilization of AI can devalue their craft and labor. Simultaneously, many writers and artists oppose the way their creative works are harvested and utilized as training materials for machine learning algorithms without their approval or payment. This renders AI particularly complicated and contentious. Nonetheless, it’s crucial to understand this backlash within a wider historical framework.

Historically, the narrative of the “death of cinema” has emerged with every significant technological transformation. For instance, the introduction of synchronized sound systems in 1926 led to the obsolescence of many silent film acting techniques, production methods, and even careers. While this caused considerable upheaval, it ultimately paved the way for new roles, such as sound technicians, and transformed rather than eliminated cinema. Similar patterns can be observed with the advent of color, television, digital media, and others. AI represents another variation of this narrative that continues to highlight shifting cultural and industry worries concerning technological influence. One of the unique aspects of cinema among other art forms is its reliance on intricate, evolving technologies. This transformation may feel unsettling, but it also presents an opportunity for all of us, including the Academy, to reassess what makes film significant.

The Academy is reportedly contemplating mandatory AI disclosure for the 2026 Oscars. Do you find this worthwhile?

BH: In general, I believe that increasing transparency is beneficial. However, as AI becomes more entrenched in production processes and workflows, overly stringent regulations could become excessively burdensome and hard to monitor. I would begin by posing the question: What do we mean by AI? Computer-generated imagery and associated algorithms have been present within the Hollywood studio system since the 1970s. When did CGI and other algorithmic tools get reclassified as AI?

I think regulations should prioritize areas where AI usage could jeopardize workers and manipulate audiences. For instance, AI-generated actors and de-aging techniques may exacerbate societal body image issues and displace real actors from their roles. Disclosure would assist the Academy and viewers in understanding what they are witnessing, thereby evaluating the ethics and better critiquing films overall.

The applications of AI in “The Brutalist” and “Emelia Peréz” are relatively minimal. How did workers perceive AI as tools to support their craft rather than supplant it?

BH: The workers did not reject AI altogether. They appeared to recognize that technological advancement is a continuous aspect of cinema and exhibited varying degrees of acceptance towards the creative potentials. They acknowledged that there are potentially beneficial applications as long as the control and decision-making regarding AI rest with them rather than studio executives pushing its adoption.

That being said, the workers seemed to view the current AI-assisted capabilities as somewhat uninspired and inadequate in enhancing (or replacing) their roles. For instance, a writer who attempted to use AI for assistance described the written outputs as “hacky” and “mundane.” Many workers presented compelling arguments for why AI could not take over the activities that authentically characterize filmmaking, such as nurturing genuine human connections on and off the screen and narrating stories that resonate with audiences.

What were your main insights from discussions with the film workers? Have those perspectives shifted as the technology has advanced over the past year and a half?

BH: Despite existing in various forms for decades, the so-called AI today exhibits a “novelty effect,” which is currently exploitable, yet likely to diminish over time. As AI becomes increasingly integrated and woven into the fabric of cinema like previous technologies, I anticipate that the anxieties surrounding it will subside.

Instead of fueling the hype cycle, we should remain composed and vigilant in pursuing ethical implementations and safeguards, as AI can inflict damages that necessitate protections for both workers and viewers. Although Hollywood unions have secured protections through collective bargaining agreements, these will require continual updates as technology evolves, and should also extend to non-union workers and those in other media fields via state and federal regulations. I would particularly advocate for policies ensuring informed consent and remuneration for artists whose works are utilized as training data for AI.

What should the public be aware of and contemplate regarding AI in filmmaking?

BH: Ultimately, it is up to us as film viewers to determine what we appreciate or dislike about AI in cinema. We wield the influence of our attention and finances to choose which films we wish to support. At the end of the day, the Hollywood studio system will invest in what is lucrative and withdraw from what isn’t. We ought to heed the voices of the workers for direction and watch films that reflect our values. Despite the prevailing concerns about AI and its current appeal, the public should keep in mind what truly renders a film valuable: the human creativity and passion behind it.

Daniela K. Rosner, a professor of human-centered design and engineering at UW, is a co-author of the journal article. This research received funding from the Labor Research and Action Network as well as the National Science Foundation.

For further inquiries, reach out to Halperin at [email protected].


Leave a Reply

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

Share This