The ChatGPT creator OpenAI has said it is working on a software called Sora that can generate minute-long videos based on text prompts.
The Microsoft-backed company said that Sora, a text-to-video model, could also take an existing still image and generate a video from it.
All About ChatGPT Creator OpenAI's Sora
Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user's prompt.
Sora can also generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background.
Sam Altman-led OpenAI said that the new tool understands not only what the user has asked for in the prompt, but also how those things exist in the physical world.
Sora has a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vibrant emotions, OpenAI said.
Sora can also create multiple shots within a single generated video that accurately persist characters and visual style.
The software follows OpenAI's ChatGPT chatbot, which was released in 2022 and created a buzz around GenAI with its ability to compose emails and write codes and poems.
OpenAI also posted the results of some prompts on X. These included a short video of an animated scene featuring a close-up of a short fluffy monster kneeling beside a melting red candle.
Another showed "a stylish woman walking down a Tokyo street filled with warm glowing neon and animated city signage".
The current model, however, has weaknesses, OpenAI said, adding that it may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect. For example, a person might take a bite out of a cookie, but afterward, the cookie may not have a bite mark.
OpenAI Working With 'Red Teamers' On Sora
OpenAI said it will take several important safety steps ahead of making Sora available.
The ChatGPT creator said it is building tools to help detect misleading content such as a detection classifier that can tell when a video was generated by Sora.
The company also said it is working with red teamers -- domain experts in areas like misinformation, hateful content, and bias -- who will be adversarially testing the new text-to-video model.
