SAN FRANCISCO — OpenAI has publicly released its new artificial intelligence video generator Sora, but the company won’t let most users depict people as it monitors for patterns of misuse.
Users of a premium version of OpenAI’s flagship product ChatGPT can now use Sora to create AI-generated videos based on written commands.
Among the highlighted examples are high-quality video clips of sumo-wrestling bears and a cat sipping coffee.
The OpenAI logo is displayed on a cellphone Dec. 8, 2023, in front of an image generated by ChatGPT’s Dall-E text-to-image model in Boston.
However, only a small set of invited testers can use Sora to make videos of humans as OpenAI works to “address concerns around misappropriation of likeness and deepfakes,” the company said in a blog post.
Text-to-video AI tools like Sora were pitched as a way to save costs in making new entertainment and marketing videos but also raised concerns about the ease with which they could impersonate real people in politics and otherwise.
OpenAI says it blocks content with nudity and a top priority is preventing the most harmful uses, including child sexual abuse material and sexual deepfakes.
The highly anticipated product received so much response after its Monday release that OpenAI temporarily paused the creation of new accounts.
“We’re currently experiencing heavy traffic and have temporarily disabled Sora account creation,” its webpage said.
OpenAI unveiled Sora this year but said it wanted to engage with artists, policymakers and others before releasing the new tool to the public.
The company, which was sued by some authors and The New York Times over its use of copyrighted works of writing to train ChatGPT, hasn’t disclosed what imagery and video sources were used to train Sora.
