OpenAI’s Sora Text-to-Video Generator Is Already Creating New Opportunities for Artists

Earlier this year, OpenAI unveiled its latest AI tool, the text-to-video generator Sora, providing access to a select group of professional filmmakers and creatives. The platform, like many such generators, is simple to use: Type in a prompt, and within a few minutes, Sora generates a soundless video up to one minute in length. As with the unveiling of the company’s flagship ChatGPT two years ago, Sora was met with both breathless media coverage and not a small measure of worry about what the technology might do to creative industries.

There was perhaps reason to worry: In June, Toys “R” Us released what it said was the first commercial made using the tool. Meanwhile, numerous other video generators have been demoed or launched by competing AI companies, including Runway, Pika, Luma Labs, and Stability AI (makers of the ubiquitous image generator Stable Diffusion). It seems a matter of when, not if, video produced by or with any of these models will become a normal part of content production for brands, advertising agencies, and beyond.

But what about artists?

In March, Alexander Reben—OpenAI’s artist-in-residence and a user of the company’s tools since 2020—posted to Instagram his first experiment with Sora, a short film of 3D sculptures rotating in space. In the caption, Reben announced that he would turn one of them, a humanoid form, into a larger-than-life marble sculpture over the next few months. Meanwhile, the Crocker Art Museum in Sacramento, California, staged “AI Am I?” a retrospective of Reben’s work, featuring numerous pieces cheekily implementing OpenAI’s generative AI technology. In one piece, a billboard refreshes with an endless supply of criticism of AI art.

While Sora has yet to be released to the public—it appears unlikely to meet its originally announced August launch date—Reben is far from the only artist exploring whether and how AI tools can produce new, meaningful work.

Sarp Kerem Yavuz, a Turkish-American artist who has employed AI in his photography for several years, suggested that artists today might do well to think of themselves more like film directors than singular creators.

“We are too quick to dismiss the kind of artistic labor that goes into AI-generated works as insufficient,” Yavuz told ARTnews. “The resulting film is comprised of everyone’s input … but we do not question the director’s artistry or authorship.”

Interestingly, Yavuz does not categorize himself as an “AI artist.” Rather, he considers AI simply another tool to work with. After generating images with a series of prompts—sometimes more than a hundred—Yavuz further manipulates and edits them. Recently, the Leslie Lohman Museum acquired his AI-generated Polaroid series “Polaroids from the Ottoman Empire,” in which he uses AI to imagine an alternate reality of queer individuals living in a contemporary Ottoman Empire that never collapsed.

Yavuz represents a growing group of artists whose use of AI does not decide their categorization. And as the boundaries become increasingly blurred, platforms like Sora could be the next step in destigmatizing tech-using artists in the industry.

Aindrea Emelife, the curator of modern and contemporary art at the soon-to-open Edo Museum of West African Art, similarly views AI as another tool at artists’ disposal.

“For art, the opportunity to explore what could be a limitless abundance of possibility is a potential source of liberation,” Emelife said. At this year’s Venice Biennale, she curated the Nigerian Pavilion, featuring works by Fatimah Tuggar that integrate augmented reality and AI to explore how colonization and globalization have eroded indigenous craft.

“Artists are innovators,” said Emelife. “We should look to them to understand the true possibilities of new technologies.”

Meanwhile, some filmmakers have argued that generative video platforms like Sora will open new possibilities for independent filmmaking outside costly Hollywood productions. Artist and filmmaker Paul Trillo has been leading that charge lately, experimenting with generative AI platforms to create short films and music videos. He received early access to Sora and has since used it to produce a short film set in the Louvre, titled Absolve, as well as Notes to My Future Self, a video art piece recently on view at the Telefónica Foundation Museum in Madrid.

For a recent music video that Trillo produced for the band Washed Out, the artist generated more than 700 clips to create a narrative of a couple’s life together through an “infinite dolly shot.” It had scenes, sets, actors, and locations that, according to the filmmaker, would otherwise have cost tens of thousands of dollars and months to produce.

Generative AI, Trillo told ARTnews, not only allows him to “revisit and resurrect some forgotten ideas, but [is also] a fundamental change in how to work.”

“There is a fluid back-and-forth process between ideation and creation. I was constantly rewriting the edit and coming up with new ideas that I wouldn’t have arrived to through any other process,” Trillo explained. “AI lends itself to the surrealism of dreams and memories, and I think it’s best suited to illustrate the subconscious. It’s the hallucinations of AI that make it uniquely suited to represent an unreliable reality. It can capture something a camera simply cannot.”

For the artist Madeline Gannon, whose work explores AI and human-robot relations, Sora offers the chance to expand AI’s reach. “I’m very excited for tools like Sora making advanced tech more accessible to artists,” Gannon told ARTnews. “It’s not just for video artists—many interactive, spatial mediums are built on top of live video feeds.” Gannon is eager to try the platform once it becomes publicly available, noting that it will allow her to explore human and robot relations in new ways.

Art made with or about artificial intelligence is far from going away. In addition to Tuggar’s work at the Nigerian Pavilion, the 60th Venice Biennale features Josèfa Ntjam’s sci-fi-style videos of AI-generated marine life in the courtyard of the Accademia di Belle Arti di Venezia. At the Pavilion of Malta, Maltese artist Matthew Attard combines AI-generated imagery with drawing, and photography.

Meanwhile, this year’s Whitney Biennial bears a theme ostensibly focused on AI, “Even Better Than the Real Thing,” though works that directly engage with the technology are few and far between.

And, of course, one of the hottest artists of the moment is Turkish-American new media artist Refik Anadol, whose AI-generated installation Unsupervised—Machine Hallucinations was acquired by MoMA in 2023 after occupying the museum’s lobby for nearly a year. Anadol has been questioning the line between human and machine creativity for close to a decade, building artworks that he told ARTnews last year he considers 50/50 “human-machine collaboration.” In his view, AI and data are merely another paintbrush for artists.

It seems that Anadol is far from the only contemporary artist to hold that view.

“I’m excited for artists to misuse AI tools in new and more interesting ways,” Yavuz said.