Artists briefly leak access to OpenAI’s Sora video generator

A gaggle of artists has leaked access to Sora, an OpenAI artificial intelligence model designed for video generation that’s currently in private alpha.

The artists made Sora’s application programming interface accessible via Hugging Face, Quartz reported today. OpenAI blocked access to the API after about three hours.

Sora debuted in February and is currently not available to the general public. It allows users to generate videos up to at least one minute in length with natural language prompts. A prompt can contain several sentences describing what objects a clip should depict, how those objects should interact and other details.

Under the hood, Sora relies on the identical transformer neural network architecture that underpins OpenAI’s language models. It uses a way called diffusion to generate clips. Diffusion produces video content through a multistep process: Sora first creates an early version of a clip that incorporates a considerable amount of noise after which steadily polishes the video in several increments. 

When OpenAI debuted the model in February, it detailed plans to share it with a limited variety of artists through an early access program. The corporate stated that the goal is to gather feedback on how Sora might be made more useful for creative professionals. OpenAI also shared the model’s API with a lot of red teamers, cybersecurity experts who concentrate on identifying vulnerabilities and other issues in AI models.

Sora’s API was posted to Hugging Face by a gaggle of about 20 artists who participated within the early access program. They explained that they leaked access to the model because they found fault in how OpenAI managed this system. “What we don’t agree with is how this artist program has been rolled out and the way the tool is shaping up ahead of a possible public release,” they wrote.

The group took issue with the undeniable fact that Sora-generated videos should be approved by OpenAI before they might be shared. Moreover, the artists criticized an initiative through which the ChatGPT developer plans to screen movies created by some early Sora testers. The initiative offers “minimal compensation which pales compared to the substantial PR and marketing value OpenAI receives,” the artists wrote on Hugging Face.

The corporate said in a press release that “a whole bunch of artists in our alpha have shaped Sora’s development, helping prioritize latest features and safeguards. Participation is voluntary, with no obligation to offer feedback or use the tool.”

OpenAI has not yet provided a release date for Sora. It did, nonetheless, share plans so as to add C2PA support within the event that it decides to make the model commercially available. C2PA is a technology that makes it easier to find out if a video was generated by AI.

If the corporate decides to commercialize Sora, it may release latest versions that address a few of the limitations within the model’s current iteration. It divulged in February that Sora sometimes struggles to “simulate the physics of a posh scene.” On occasions, the model also misinterprets prompts that include “spatial details” comparable to the direction through which an object should move. 

OpenAI’s DALL-E series of image generation models, which is already commercially available, has received several major upgrades since its release. The corporate has also built it into ChatGPT. It’s possible OpenAI will add an analogous integration to Sora if and when it decides to make the video generator commercially available. 

Image: OpenAI