A dramatic leak of OpenAI's highly anticipated Sora video generation model sent shockwaves through the AI community this week.
But the real story isn't just about the technology—it's about growing tensions between AI companies and the creative professionals they rely on.
What happened? And what does it mean for AI?
I got the scoop from Marketing AI Institute founder and CEO Paul Roetzer on Episode 125 of The Artificial Intelligence Show.
On Tuesday, November 26th, a group calling themselves the "Sora PR Puppets" briefly gave the public unauthorized access to Sora's API through the AI platform Hugging Face.
Users were able to generate 10-second videos in 1080p resolution simply by typing text descriptions before access was revoked hours later.
And it seems this is legit.
“It does seem like it’s real,” says Roetzer. “The videos we’re seeing, the leaks we’re seeing are real.”
But here's what makes this leak even more fascinating:
The group wasn't protesting the technology itself. Instead, they were protesting how OpenAI treats its early testers and creative partners.
The Sora PR Puppets explained their motives in a short statement accompanying the leak, writing:
“We received access to Sora with the promise to be early testers, red teamers and creative partners. However, we believe instead we are being lured into ‘art washing’ to tell the world that Sora is a useful tool for artists.”
Their main issues appear to be two-fold:
The group took pains to note that they’re not against using AI to generate video—or any other artistic medium, writing:
“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). What we don't agree with is how this artist program has been rolled out and how the tool is shaping up ahead of a possible public release.”
OpenAI pushed back on these claims, according to TechCrunch.
The company emphasized that Sora remains in a "research preview" phase and that artist participation is entirely voluntary, with support provided through grants and events.
The leak also gave us an early look at Sora’s current capabilities.
One prominent AI researcher, Ethan Mollick, noted that the leaked videos appeared to be "an order of magnitude better than competing tools."
But he was quick to caution that we don’t know if the leaked footage represents Sora’s true capabilities out of the box.
And, while Sora is impressive, it’s still very early, warns Roetzer. The technology still has plenty of barriers.
"I think people see these demos and think, oh my gosh, we're there. We're going to be able to create one minute long videos with a simple text prompt," he says.
"But it's just not where the tech seems to be. There are a lot of barriers to this working and consistently generating those kinds of outputs."
He still finds today’s AI video tools limited, even though they have made significant progress in the past couple years.
“I think you still have to have deep expertise in video production to get value from these tools or you have to spend a whole bunch of time working with them.”
So, the jury’s still out as to whether or not Sora can deliver on all the hype generated by the example videos created during the leak.
For an answer to that, we’ll need to wait for OpenAI to actually release the tool.