He Helped Directors Imagine Their Films. Then AI Let Him Make One Himself

Robert Gaudette makes most of his living drawing pictures of films other people will shoot.

Directors and agencies hire him to visualize treatments, pitch decks and storyboards with generative AI. That work, Gaudette says in two video interviews with me this month, “still is ninety-five percent of what I do today.”

Then ByteDance’s Seedance video model gave him a reason to test the other end of the production chain. Could the person making images that sold a production also make the finished pictures?

Gaudette cast, shot, scored and cut a story he had already written, working alone at his desk over about two weeks.

“My short film was actually my effort to take a story that I had written and go through the entire production process,” he says.

The demonstration got away from him.

A Face Only a Mother Could Love, a seven-minute, 50-second story about a disfigured Parisian who dances alone in his apartment, was entered into festivals “on a whim.” It won the Grand Prix at Runway’s AI Festival in New York, carrying a $50,000 prize and 1 million Runway credits, then won the Reply AI Film Festival in Venice and best picture at the OMNI Hyperphantasia festival in Sydney.

Gaudette works out of Toronto. He spent about 15 years in technology, a decade as a commercial and portrait photographer and eight and a half years in community and employment work before moving into AI pre-production.

The filmmaking arrived suddenly. The constituent parts did not.

Other AI filmmakers who reached Hollywood’s attention this year arrived with something already working in their favor.

Gaudette was not using AI film to enter the industry, but to move further into one where clients already knew his work.

He Wanted To Move Beyond Previs

Previs helps someone else make the production. Gaudette wanted to find out whether he could deliver the production itself.

Pre-production is forgiving territory for generative AI. “Everyone seems very amenable about using AI in pre-visualization,” Gaudette says.