Maggie Gyllenhaal, the president of the Venice jury, recently shared her experience experimenting with artificial intelligence (AI) while working on her short film “Flesh Impact,” which features Dakota Johnson as a young Marilyn Monroe. The project was initially inspired by a blend of nostalgia and modern technology, but Gyllenhaal ultimately chose to abandon the AI component, noting it detracted from the emotional richness of Johnson’s performance.
A Journey into AI and Artistry
During a symposium that included other industry veterans like composer Alexandre Desplat, filmmakers Lee Daniels and Shannon Murphy, producer Lorenzo Mieli, and artist Marc Quinn, Gyllenhaal explained the film’s unique narrative structure. “Flesh Impact” portrays a 100-year-old Marilyn Monroe alongside flashbacks to her 30s, with Johnson taking on the challenging role.
Initially, there was a push from those who commissioned the project to incorporate AI. “They weren’t using OpenAI; they were using licensable AI, training it off of Marilyn Monroe films that we had licensed,” Gyllenhaal recounted. Intrigued and unsure of what the outcome would be, she embraced the challenge. “I went for it. I was like, all right, I’m curious, I don’t know, let’s see,” she said.
The Human Element
However, after extensive efforts to digitally transform Johnson into Monroe, Gyllenhaal realized that despite the technical work, the AI-generated version failed to resonate as powerfully as Johnson’s authentic portrayal. “In the end, I threw it away because what was coming through Dakota playing Marilyn Monroe was so much more moving, human, alive,” she revealed.
Gyllenhaal voiced her disappointment with the lack of emotional depth in the AI representation, stating, “It fundamentally did not work. All the reasons we decided to make this project to begin with were killed off in the AI version.”
Concerns About AI in Filmmaking
Beyond the creative setbacks, Gyllenhaal also expressed concerns about the environmental implications of AI technology. “There’s a big conversation in South Memphis about the data centers. That’s a problem, so we should acknowledge that,” she noted, highlighting the broader issues surrounding AI development.
Adding to the discussion, she referenced Geoffrey Hinton, a pioneer in AI research, who has raised concerns regarding the technology’s future. “He said that he thought we should program AI with a maternal instinct,” she shared, reflecting on the fascinating implications of such a notion.
The Financial Dilemmas of AI
Gyllenhaal highlighted skepticism regarding the much-touted “democratization” that AI is believed to bring to the creative process, questioning the motivations behind such claims. “It concerns me a little bit because they own it. Who’s making the money off it? The people that are pitching that,” she quipped, drawing appreciative applause from the audience.
As Gyllenhaal delved deeper into the topic, she warned that filmmakers might suffer if AI technology becomes more prevalent in commercial cinema. She cited a stark prediction: “$40 million will now be able to cost $4 million if you do it with AI… and maybe even eventually $40,000.” While this could open doors for emerging creators, she cautioned, “If you can make a commercial good enough movie for that price, who’s gonna give you $10 million to do it the old way?”
In wrapping up her reflections, Gyllenhaal underscored the importance of authentic human storytelling in film, reiterating that algorithms cannot replace the essence of what makes performances truly impactful.
