SCA students, faculty and staff have also been hearing from a wide spectrum of speakers, and storytellers, about what the future could hold for scholars and creators in a new world of creativity, where AI is integrated into every aspect of development, production and distribution. New to the schedule this fall were two new undergraduate courses. “AI and Creativity,” offered an introduction to “the history, ethics and practices of generative AI in the context of the cinematic arts.” Among the assignments was creating a music video using text-to-video tools. The other class, “Contemporary Directing Practice” asked “What is directing when you’re working with text-to-image and text-to-video tools?” Holly Willis, Chair of Media Arts + Practice division, who has become a leading voice on the use of AI tools, says “AI is providing artists with a new visual vocabulary and, combined with virtual production, a new workflow.”

In March, Willis was named co-director of USC’s new Center for Generative AI and Society. With $10 million in funding, the Center will convene experts and explore the transformative impact of artificial intelligence on culture, education and media across the 22 USC schools. Working with co-director Bill Swartout, research professor at the USC Viterbi School of Engineering and CTO for USC Viterbi’s Institute for Creative Technologies, Willis will fulfil university president Carol Folt’s mandate to chart a course of vigorous exploration of “the intersection of ethics and the use and evolution of generative AI.” In that capacity Willis — an author, artist and 17-year SCA veteran will be “raising critical awareness, contextual understanding and basic AI literacy across SCA while also learning from our community what's important to them with regard to intellectual property, notions of human creativity, and more.”

For the SCA community, the important tests will happen in the creating, and the discussions that the projects elicit. Another collaboration between the School of Cinematic Arts and the Annenberg School for Communication & Journalism has allocated twenty spark grants to fund faculty projects on the creative use of AI. An upcoming exhibit will feature work created in the AI classes, along with discussions about ethics, creativity, and inevitably the kinds of jobs graduating students can compete for if they learn to master these tools. For example, “Prompt engineer” is a new entertainment job description. “Learning how to dialogue with the AI is part of the process, translating what you want into something the algorithm can figure out,” explains Willis. “Photoreal” is not a recommended prompt, because it’s not a word typically used to describe iconic visuals, and LLMs are trained by associating words with pictures. “Cinematic” or “photograph of” tend to produce better results. The models can even translate lens traits, like aperture. Midjourney’s “chaos parameter,” a value between 1 and 100, controls the randomness of an image. Imagination, quantified.

To Holly Willis, artificially generated images are “metaphors that capture the cultural transition from an emphasis on the photographic to the data image,” while Yves Bergquist is more utilitarian, informed by the ETC’s perch at the intersection of academe and industry. “The technology is only worthwhile insofar as it is empowering creativity,” he says. “Sure, SCA makes sure the students understand what’s powering the models, but the only thing that matters is the product that results.” AI’s influence will be more additive than displacement, he believes, citing “micro-workflows.” “It will automate things like neural radiance fields.” Also known as NeRFs, neural radiance fields transform 2D into 3D images. “That will definitely accelerate virtual production,” Bergquist says. “Other micro workflows — language translation and that sort of thing — will be optimized. But don’t expect push-a-button-get-a-movie kind of stuff.”