OpenAI has made its new AI video generator, Sora, publicly available, but the company restricts most users from creating videos of people to monitor for potential misuse.
Subscribers of the premium version of OpenAI's main product, ChatGPT, can now utilise Sora to generate AI-generated videos instantly from text prompts. Featured examples include high-quality clips showing bears engaged in sumo wrestling and a cat enjoying a cup of coffee.
However, only a select group of invited testers has the ability to produce videos featuring humans, as OpenAI aims to address concerns regarding the misuse of likenesses and deepfakes, according to a statement in a blog post.
Text-to-video AI solutions like Sora have been promoted as a cost-effective method for producing new entertainment and marketing videos, yet they have also raised alarms about the potential for impersonating actual individuals in political contexts and other areas.
OpenAI claims it is restricting content that contains nudity and prioritising the prevention of the most harmful applications, such as child sexual abuse material and sexual deepfakes.
The creators revealed they have faced high demand and have paused account creation temporarily. “We're currently experiencing heavy traffic and have temporarily disabled Sora account creation," states its webpage.
OpenAI initially introduced Sora earlier this year, but the company expressed a desire to first engage with artists, policymakers, and others before making the new tool accessible to the public.
The organisation, which has faced lawsuits from certain authors and The New York Times regarding its use of copyrighted text to train ChatGPT, has not revealed which imagery and video resources were utilised to train Sora.
Follow us
