A brief public opening into OpenAI’s Sora video generator has put fresh attention on the company’s relationship with artists involved in early testing. The access appeared through a Hugging Face project connected to the unreleased Sora API, and it was framed by the group behind it as a protest against how OpenAI is running the program.
The episode was short-lived, but it surfaced several pressure points at once: unpaid creative work, control over what artists can share, safety claims around a research preview, and the competitive race to build convincing AI video tools.
What Appeared Online
On Tuesday, a group published a project on Hugging Face that seemed to connect to OpenAI’s Sora API, which is not publicly available. The group used authentication tokens, apparently from an early access system, to create a front end that allowed users to generate videos with Sora.
Through that front end, users could type a short text description and create 10-second videos up to 1080p resolution. TechCrunch reported that the queue was long when it tried the tool, while several users on X posted samples. Many of those samples carried OpenAI’s distinctive visual watermark.
The project did not stay usable for long. As of 12:01 p.m. Eastern, the front end was no longer working. TechCrunch suggested that OpenAI and/or Hugging Face may have revoked access. The group later claimed that, after three hours, OpenAI temporarily shut down Sora’s early access for all artists.
Outside observers also pointed to technical signs that the access was real. Tibor Blaho wrote on X that the project was using the OpenAI Sora API endpoint to generate and download videos, with hardcoded request headers and cookies from the Hugging Face space environment config.
Why The Group Says It Acted
The group called itself “Sora PR Puppets.” Its complaint was not that AI video should never be used in art. Instead, it argued that OpenAI’s artist program had become too focused on public relations and not enough on creative critique or fair treatment of contributors.
In a post attached to the front end, the group wrote: “Hundreds of artists provide unpaid labor through bug testing, feedback and experimental work for the [Sora early access] program for a $150B valued [sic] company.” It added: “This early access program appears to be less about creative expression and critique, and more about PR and advertisement.”
The group initially did not identify its members. Later in the day, it began listing some names in the Hugging Face attachment and in a separate petition.
The core accusation was that OpenAI was benefiting from artists’ testing, feedback and experimental work while keeping the public view of Sora tightly managed. According to the group, early testers, including red teamers and creative partners, were being pressured to help produce a positive story around Sora.
The Approval Dispute
A major part of the group’s complaint involved control over Sora outputs. The group claimed that every Sora output must be approved by OpenAI before it can be shared widely. It also said only a few creators in the program would be selected to have Sora-created works screened.
That claim matters because early access programs can shape how a tool is understood before release. If the public mostly sees approved examples, the tool’s weaknesses may be less visible. The group argued that this created a misleading picture of Sora’s capabilities.
The group also wrote: “We are not against the use of AI technology as a tool for the arts (if we were, we probably wouldn’t have been invited to this program).” It continued: “What we don’t agree with is how this arti st program has been roll ed out and how the tool is shaping up ahead of a possible public release. We are sharing this to the world in the hopes that OpenAI becomes more open, more artist friendly and supports the arts beyond PR stunts.”
Those statements positioned the leak as a protest over process, not a rejection of AI video generation itself. The group’s message focused on compensation, transparency and the boundaries placed on participating artists.
OpenAI’s Response
OpenAI described Sora as still being in a “research preview.” A spokesperson said the company is “working to balance creativity with robust safety measures for broader use.”
The spokesperson also said: “Hundreds of artists in our alpha have shaped Sora’s development, helping prioritize new features and safeguards.” OpenAI said participation is voluntary and that artists have no obligation to provide feedback or use the tool.
OpenAI said it had offered artists free access and would continue supporting them through grants, events and other programs. The company also said it believes AI can be a powerful creative tool and is committed to making Sora useful and safe.
The spokesperson added that artists “have no obligations” beyond “responsibly” using Sora and not sharing confidential details while Sora is under development. However, the spokesperson did not clarify what “responsible” use means or which details OpenAI considers confidential.
Why Sora Is Under Pressure
The leak landed at a difficult moment for Sora. Since its debut earlier this year, the video generator has faced technical setbacks while rivals in video generation have moved aggressively.
One of Sora’s co-leads, Tim Brooks, left OpenAI for Google in early October. In a recent Reddit AMA, OpenAI chief product officer Kevin Weil said Sora was being held back by the “need to perfect the model, get safety/impersonation/other things right, and scale compute.”
Technical limits have also been part of the story. Per The Information, the original system, revealed in February, took more than 10 minutes of processing time to create a one-minute video clip. Consistency was another challenge. Filmmaker Patrick Cederberg had to generate hundreds of clips before getting a usable one, because the model struggled to maintain styles, objects and characters across videos.
The leaked Sora appeared to be a faster, “turbo” variant, according to code uncovered by X users. The code also hinted at style controls and limited customization options. Per The Information, OpenAI has been training Sora on millions of hours of high-quality clips across a range of styles and subjects to improve generated video quality.
At the same time, competitors have been making visible moves. In September, Runway signed a deal with Lionsgate, the studio behind “John Wick,” to train a custom video model on Lionsgate’s movie catalog. Roughly a week later, Stability recruited “Avatar” director James Cameron to its board.
OpenAI was said to be meeting with filmmakers and Hollywood studios earlier this year to demo Sora, and ex-CTO Mira Murati attended Cannes. But the company has not announced a collaboration with a major production house.
The Sora leak, then, was more than a temporary access problem. It exposed a dispute over who gets to test powerful AI video systems, how their work is valued, and how much control a company should have over the first public image of a tool still under development.