OpenAI shows off first examples of third-party creators using Sora

OpenAI has been releasing a steady drip of videos generated by its new highly realistic AI model, Sora, but the text-to-video (and image-to-video) tool remains out of reach to the public still.

Now, for the first time, the generative AI startup is showing off Sora creations made by a select group of outside filmmakers, artists, advertising agencies, and musicians that have been given access to the model.

“While we have many improvements to make to Sora, we’re already getting a glimpse of how the model can help creatives bring ideas to reality,” the company wrote in a blog post published today including 7 of the videos generated by its select group of Sora invitees. Among those whose work is featured include:

1. Walter Woodman, Sidney Leeder, Patrick Cederberg – Members of shy kids, a multimedia production company based in Toronto. Walter directed the short film “Air Head.”
2. Paul Trillo – Multi-disciplinary artist, writer, and director.
3. Nik Kleverov – Creative Director and Co-Founder of Native Foreign, an Emmy-nominated creative agency.
4. August Kamp – Musician, researcher, creative activist, and multidisciplinary artist.
5. Josephine Miller – Co-Founder and Creative Director of Oraar Studio, specializing in 3D visuals, augmented reality, and digital fashion.
6. Don Allen Stevenson III – Digital AR/XR Artist, speaker, and consultant.
7. Alex Reben – Sculptor/Artist and OpenAI’s Artist In Residence.

Here’s one of the videos from Paul Trillo:

Naturally, they all sing its praises — why else would OpenAI publish their work if they didn’t like using the tech? As a promotional exercise, it makes total sense.

Still, these Sora videos show the potential of the technology beyond the walls of OpenAI. The videos also come following a report by Bloomberg that OpenAI executives have met with Hollywood filmmakers and studios to drum up interest in using Sora for conventional, big budget filmmaking.

OpenAI’s Sora publicity campaign also comes at a time when AI video and generative AI for content creation more generally is being met with more open skepticism and derision from artists and viewers.

Just last week, the new indie horror film Late Night With The Devil was criticized by dozens online for its use of AI to create still images for certain transitions and, seemingly, set decor — with some even calling for a boycott on it and any films that used AI instead of hiring artists using more traditional methods.

Today, in reaction to the new wave of third-party Sora videos, former Stability AI executive and current Fairly Trained CEO Ed-Newton Rex posted on X that OpenAI was “Artistwashing: when you solicit positive comments about your generative AI model from a handful of creators, while training on people’s work without permission/payment.”

Fairly Trained is a new non-profit that certifies AI models that can demonstrate they have been on only licensed or public domain data. So far, OpenAI has kept mum about what specific data it used to train Sora, with chief technology officer Mira Murtai telling The Wall Street Journal in a widely-circulated (and mocked) video interview that the company used “publicly available and licensed data,” later clarifying it used at least some videos from its license agreement with Shutterstock.