ChatGPT maker OpenAI halted the rollout of its new text-to-video generator, Sora, after testers leaked the tool to the public over concerns about exploitative business practices.
Artists who were given early access leaked Sora on artificial intelligence platform Hugging Face under the username PR-Puppets. Just three hours after its limited release, OpenAI removed access to Sora for all artists.
The artists said in an open letter Tuesday, addressed to “Corporate AI Overlords,” that they were being “lured into ‘art washing’ to tell the world that Sora is a useful tool for artists.” They said that while they are not against AI being used in creative work, they don’t agree with how the artist program was rolled out and how the tool is shaping up.
“We are not your: free bug testers, PR puppets, training data, validation tokens,” the artists wrote.
As of Wednesday morning, more than 630 people had signed the letter. The situation highlights both the ethical and existential tension between the art world and AI.
An OpenAI spokesperson said that it is still focused on balancing creativity with safety measures before rolling out the tool for broader use.
“Hundreds of artists in our alpha have shaped Sora’s development, helping prioritize new features and safeguards,” a spokesperson said. “Participation is voluntary, with no obligation to provide feedback or use the tool. We’ve been excited to offer these artists free access and will continue supporting them through grants, events, and other programs.”
Beta testers, including filmmakers and visual artists, behind the open letter criticized the early access program for being “less about creative expression and critique, and more about PR and advertisement,” noting that “every output needs to be approved by the OpenAI team before sharing.”
Access to videos generated using Sora was removed, but some anonymous users on Hugging Face archived dozens of videos and corresponding prompts and shared them on other social media platforms, including X (former Twitter) and Reddit.
In May, OpenAI said that it would provide access to Sora to “red teamers to assess critical areas for harms or risks,” as well as to “visual artists, designers, and filmmakers to gain feedback on how to advance the model to be most helpful for creative professionals.”
Sora was first released to testers in February. The tool is what’s known as a diffusion model, meaning it generates video by starting with a blurry, static-filled image that smoothes it into a polished final version.
