In the third episode of “Creative Dialogues,” an interview series produced by the film division of generative AI startup Runway, multimedia artist Claire Hentschker expresses her fear that AI will commodify the artistic process to the point where art becomes homogenized, regressing to a kind of derived equality.
“Are you getting this increasingly narrow average of existing things?” she asks. “And then as that is averaged out, everything will just be a mass?”
Those are the questions I kept asking myself Wednesday at a presentation of the top 10 finalists at Runway’s second annual AI Film Festival, who are available on demand on the Runway website starting this morning.
Runway held two premieres this year, one in Los Angeles and one in New York. I attended the one in New York, which took place at Metrograph, a theater known for its art and avant-garde works.
I’m happy to report that AI is not accelerating into a bubble future… at least not yet. But a skilled eye as a director—the human touch—makes a clear difference in the effectiveness of an “AI movie.”
All films submitted to the festival incorporated AI in some form, including AI-generated backgrounds and animations, synthetic voiceovers, and bullet-style special effects. None of the elements seemed to be on par with what more modern tools offer. Sora from OpenAI can produce, but that was to be expected, given that most of the presentations were finalized at the beginning of the year.
In fact, it used to be obvious (sometimes painfully so) which parts of movies were the product of an AI model, not an actor, cameraman or animator. Even the otherwise solid scripts were sometimes let down by the disappointing AI generative effects.
Take, for example, “Dear Mom” by Johans Saldana Guadalupe and Katie Luo, which tells the story of a daughter’s loving relationship with her mother, in the daughter’s own words. It’s a tear. But a scene on a Los Angeles freeway with all the telltale weirdness of AI-generated videos (e.g. deformed cars, strange physics) broke the spell for me.
The limitations of current AI tools seemed to hold back some movies.
like my colleague Devin Coldewey recently wrote, control with generative models, particularly those that generate videos, is difficult to achieve. Simple issues in traditional cinema, such as choosing the color of a character’s clothing, require solutions because each shot is created independently of the others. Sometimes even alternative solutions don’t work.
The resulting disjointedness was evident at the festival, where several of the films were little more than tangentially related vignettes linked by a narrative and a soundtrack. “L’éveil à la création” by Carlo De Togni and Elena Sparacino demonstrated how boring this formula can be, with slideshow-like transitions that would make an interactive storybook better than a movie.
“Where do grandmothers go when they get lost?” by Léo Cannone? He, too, falls into the vignette category, but triumphs nonetheless thanks to a heartfelt script (a boy describing what happens to his grandmothers after he passes away) and an exceptionally strong performance from his child star. . The rest of the audience seemed to agree; The film received some of the liveliest applause of the night.
And for me, that really sums up the festival in a nutshell. Human contributions, not AI, often make the difference. The emotion in the voice of a child actor? That stays with you. AI generated funds? Less.
This was certainly true for the festival’s Grand Prize winner “Get Me Out,” which documents a Japanese man’s struggle to recover from the psychological toll of his immigration to the United States as a child. Filmmaker Daniel Antebi depicts the man’s panic attacks with the help of AI-generated graphics, graphics that I ultimately found less successful than cinematography. The film ends with a shot of the man walking across a bridge just as the streetlights dotting the pedestrian lane flicker one by one. He’s haunting, and beautiful, and it sure took a long time to capture him like that.
It’s entirely possible that one day generative AI will be able to recreate scenes like this. Perhaps cinematography will eventually be replaced by prompts, a victim of ever-growing data sets (albeit with Worrying copyright status) in which startups like Runway and OpenAI are training their video generation models.
But that day is not today.
When the screening ended and the honorees marched to the front of the theater to take a photo, I couldn’t help but notice the cameraman in the corner documenting the whole thing. Perhaps, on the contrary, AI will never replace some things, like the humanity that we humans deeply long for.