OpenAI's new text-to-image model Sora is still in its early stages, but the first artists and creatives are already testing it. OpenAI has now published some impressions of the creative use of Sora.
Since the limited launch of Sora a month ago, OpenAI has been working with visual artists, designers, creative directors and filmmakers to find out how Sora can be integrated into their creative processes. The company is now sharing some early experiences with Sora AI on its blog.
"Sora is at its most powerful when you’re not replicating the old but bringing to life new and impossible ideas we would have otherwise never had the opportunity to see," says director and writer Paul Trillo.
This is the motto of the first demos. Toronto multimedia production company shy kids used Sora for their short film "Air Head," about a balloon man.
"We now have the ability to expand on stories we once thought impossible," say Walter Woodman, Sidney Leeder and Patrick Cederberg. Sora is particularly good at creating completely surreal things, says director Woodman.
Nik Kleverov of creative agency Native Foreign sees the technology as a way to quickly visualize and iterate creative concepts for brand partners without budget constraints limiting the creative process.
Josephine Miller, Creative Director at Oraar Studio, expresses that Sora has empowered her by allowing her to "bring to life ideas" she held for years, which were previously technically unachievable. The ability to swiftly conceptualize has been driving her creative process and aiding in the development of her storytelling skills, she explains. "It's enabling me to translate my imagination with fewer technical constraints," she adds.
Artists like August Kamp and Don Allen Stevenson III also report that Sora opens up new possibilities for realizing their visions. AR/XR artist Don Allen III appreciates Sora's "freedom" from physical laws and conventions. OpenAI Artist in Residence Alexander Reben uses Sora as a starting point for his 3D sculptures and explores the possibilities of photogrammetry.
You can find many more demos on the OpenAI blog. The Sora video AI model will be released later this year. Details on pricing and generation time are not yet available.