Defunct Google AI Model Veo 1 Inspires San Francisco Exhibition

(MENAFN- USA Art News) A new exhibition at tiat gallery in San Francisco, titled“Post Generation,” features work by three artists who drew inspiration from an archive of videos created by Veo 1, a defunct Google artificial intelligence model developed in 2024. This initiative explores the impact of AI's rapid development cycle on artistic practice.

Originally released for internal use, Veo 1 produced six-second videos with prolonged generation times and often surreal results, such as a six-legged horse or a three-limbed hominid. Despite its limitations, Google researchers generated hundreds of thousands of videos before the model was superseded by publicly released versions Veo 2 and Veo 3. Due to the resource-intensive nature of these models and the efficiency of newer versions, older models like Veo 1 are typically retired, and their outputs are not publicly accessible.

Tina Tarighian, a senior creative technologist at Google's Creative Lab and the exhibition's organizer, initiated the project.“Because these models are so resource intensive and because each new model is faster and more efficient, it's hard to justify keeping old models around. No one can generate using Veo 1, and the outputs it produced aren't publicly available,” Tarighian stated.“So I thought, what can we do with these leftover scraps of a model? Can we gift this archive to artists?”

Tarighian shared a portion of this archive with choreographer Zelia ZZ Tan, creative engineer Karyn Nakamura, and fiber artist Haley Mortin, commissioning them to create new works. The resulting exhibition serves as a multimedia examination of this technology and a historical record of a specific moment in its evolution.

For Mortin, this project aligned with her existing practice of incorporating AI-generated images and research artifacts into knit works. Her interest in these artifacts stems partly from her previous role as a data annotator, where she categorized and tagged images to train AI models.“It was 2018, I had just graduated college, I didn't know what AI was, and I needed money,” Mortin recalled.“I thought it was just a data entry job.” Over two years, she labeled images for objects, subjects, and even subjective emotions.

Mortin focused on the unconventional text forms generated by Veo 1, which she described as“two flavors of nonsense, these glyph-like shapes or empty phrases.” She noted that visual AI models interpret text as texture, unlike large language models. This observation drew parallels to 18th-century embroidery samplers, where women learned writing by stitching text and often used these works to document family histories. Mortin's piece, Soup Is Made of Water (10 Samplers) ( 2026 ), features ten knitted works depicting various garbled texts from Veo 1, reflecting on the historical production and preservation of text. Rendered in black and white yarn, these pieces include a looping scribble, a calligraphic attempt to spell“beauty” [b eurt], and a sigil tattoo design.

Nakamura explored the research histories behind generative models by combining several video technologies in her work, Labor of Light ( 2016 ). This piece functions as an“altar to the history of video production.” A white device, suspended by cables, projects a beam of light that appears steady but contains the analog signal of Quicker 'n a Wink ( 1940 ), a film detailing MIT's Harold Edgerton's development of high-speed photography. This light leads to a projector casting a collage of Veo 1 videos onto a white cloth, emerging from vibrant RGB noise, a core element of generative AI. Beneath the cloth, toothpicks secure printouts of pivotal research papers that advanced modern AI generation techniques, including diffusion.

“I find the process of image making and the process of video and video systems incredibly beautiful, there's a lot of craft in there,” Nakamura commented. She expressed concern that the rapid pace of development in AI makes its internal workings increasingly opaque, limiting creative intervention. Her work often highlights the“glitches and gaps” of earlier models like Stable Diffusion 1.5. Nakamura reflected,“Because I like Stable Diffusion 1.5, sometimes I think, 'Oh, am I being a Luddite?' Like, what is even progress right now?”

Choreographer Zelia ZZ Tan addresses this tension in a ten-minute dance performed at the exhibition's opening on October 8. The performance, which integrates choreography with animated avatars, captures the movements and gestures observed in Veo 1 videos.“I searched the archive through movement rather than only through words when I started to realize that between generated bodies, physical bodies, and the computational body, there are a lot of similarities and also a lot of rules based on how they can perform,” Tan explained. She emphasized the vulnerability involved in dancing with the constraints of a generated body.“It's such an honor to have a body, to be able to be present there, and maybe I'm not showing the most trained technique, but to listen, to be hesitant on stage, to wait for something to happen, that's part of the power of being there as a performer.”

Source: ARTnews

MENAFN12102026005694012507ID1111796458