On February 15, OpenAI unveiled Sora, a new model capable of transforming text into a realistic video. The promise is incredible, but Sora is not yet accessible to the general public.
In a few years, will we be able to easily distinguish a video filmed on a smartphone and a video generated by Sora, OpenAI's new artificial intelligence? After causing a technological revolution with ChatGPT, Sam Altman's company could well be about to do it again with Sora, which is capable of generating ultra-realistic videos using prompts. OpenAI's ambitions are limitless.
Can we try Sora?
Generally, OpenAI deploys its new features in two stages. The company lets its paid ChatGPT Plus subscribers test them first, then offers them to everyone using the service. With Sora, the method is different.
Since OpenAI knows the dangers potentially caused by such technological advancement (anyone could generate false information authenticated by a video), Sora is exclusively reserved for Red members, a community of experts hand-picked to assess risks. OpenAI also invited a few VIPs, but does not intend to let the general public get their hands on the AI for the moment. “ We will secure access to a few artists, designers and filmmakers to get feedback on how to move this model in the best direction for creative professionals » indicates the company on its site.
Ironically, the only way to try Sora today is to send a prompt… to Sam Altman. Proud of the project, the boss of OpenAI offers Internet users to send him instructions, to generate videos. The results are often impressive, although one can imagine that Sam Altman filters the results.
Eventually, of course, Sora will most likely be available on its own site or directly in ChatGPT.
While waiting to be able to generate videos in ChatGPT, it remains possible to design images with Dall-e, the other OpenAI model. Sora also has some competitors, like Google Light.