AI is restricted to its training data, and that it must not be able to imagine the world the way humans would. That, is a past tense.
And that, is also an understatement, because since generative AI comes into play, the world is in the eyes of computers, is becoming a simulator. And as time advances, that simulator becomes increasingly difficult to tell apart from reality.
Since OpenAI introduced DALL·E, and later DALL·E 2 and DALL·E 3, the company is trying to give AI the ability to imagine the world without having to even experience it.
Then, the company went a step further, when it introduced Sora.
Soon after, OpenAI reaffirmed its position as a trailblazer, leaving pretty much everyone else behind its business dust.
But fast forward, as others are playing catching up, OpenAI seemingly live in the past.
Fast forward, however, and the landscape has shifted. While others have launched their own text-to-video tools comparable to Sora, OpenAI has kept Sora in beta, seemingly stuck in the past.
And this time, testers are annoyed.
The AI community is in uproar over the unauthorized leak of Sora, shared on Hugging Face by a user called "PR-Puppets."
The open letter brings light to critical issues around innovation, labor practices, and corporate accountability.
" [...] we are not your: free bug testers, PR puppets, training data, validation tokens."
"Hundreds of artists provide unpaid labor through bug testing, feedback and experimental work for the program for a $150B valued company. While hundreds contribute for free, a select few will be chosen through a competition to have their Sora-created films screened — offering minimal compensation which pales in comparison to the substantial PR and marketing value OpenAI receives."
" [...] Furthermore, every output needs to be approved by the OpenAI team before sharing. This early access program appears to be less about creative expression and critique, and more about PR and advertisement."
The disgrunted group of artists said that they're releasing (leaking) this AI tool to give everyone else the taste of "what ~300 artists were offered: a free and unlimited access to this tool."
They also object to OpenAI’s content approval requirements for Sora, which apparently state that "every output needs to be approved by the OpenAI team before sharing."
The group hopes that the leak would "denormalize" proprietary AI technology from billion dollar brands that have exploited artists by not paying or compensating them.
Sora, still considered a cutting-edge text-to-video generator known for creating high-quality videos from mere text input.
It's still considered by many as the milestone in the advancement of generative AI, in which the diffusion model can deliver precise text-to-visual alignment through its temporal coherence.
OpenAI envisions Sora as a key milestone in the pursuit of Artificial General Intelligence (AGI).
While the technology still faces limitations, including difficulties in replicating complex physics and ensuring content safety, which are areas that require further refinement, the Hugging Face discussion platform sees Sora as "a mesmerizing display of technical prowess."
The model’s ability to produce "visually coherent narratives" in video form has been praised as a landmark achievement in generative AI.
This leak has since sparked debate over ethics, technological progress, and artistic rights.
Responding to the situation, OpenAI said that Sora is still in research preview, adding that the company is still working to balance creativity with strong safety measures for broader use.
"Hundreds of artists in our alpha have shaped Sora’s development, helping prioritize new features and safeguards. Participation is voluntary, with no obligation to provide feedback or use the tool," a spokesperson for the company said.
PR Puppets then updated its open letter, to name their group as "sora-alpha-artists."
The group also shared a public petition, inviting others to support the message in their open letter.
OpenAI responded by pausing the development of the project.
Despite this, users who managed to have access to the interface before it was removed were able to generate 10-second video clips in up to 1080p resolution, and showcasing them to the internet.
These clips were created using what seems to be an optimized "turbo" version of the model initially demonstrated by OpenAI in February.
As for the delay, previous OpenAI CTO, Mira Murati, said in March that the company planned to release Sora publicly by the end of 2024. However, CPO Kevin Weil later said that the platform's deployment has been delayed by the "need to perfect the model, need to get safety/impersonation/other things right, and need to scale compute!"