openAI announces Sora to convert text to video
OpenAI has officially announced its new artificial intelligence model, Sora, for converting texts into videos, which the company says can create video of real and imagined scenes from textual descriptions, including videos that include complex scenes and different types of motion.
And he said Sam Altman The CEO of OpenAI said in a post on the X website that the Sora AI model is now available to a select group of users, including the team that is evaluating the potential risks of using the new model, in addition to a group of content producers, filmmakers, and designers to obtain their feedback on the model.
What is Sora from OpenAI?
It provides a new artificial intelligence model of OpenAI company Which it calls Sora, a tool for users to convert text into video, so that the user only has to describe any scene in text to get a high-quality video of up to one minute, and the company that develops GPT chat said that it is working to teach artificial intelligence how to understand and simulate The moving physical world.
OpenAI confirmed that its new artificial intelligence model, Sora, is capable of creating videos of complex scenes with multiple characters, with specific types of movement and fine details of the scene and background.
The company says that the model is not only able to understand the user's text prompts to create the video, but it is able to understand how the things described within the video exist or relate to the real world. Sora can also create multiple shots in a single video that are created precisely with characters and commitment. With a certain visual style.
But the new model of artificial intelligence for converting texts to video, Sora, still has some shortcomings, according to the developer, as it may sometimes find difficulty in accurately simulating the natural movement of a scene, and it may not understand in specific cases the relationship between cause and effect, such as showing a video scene to a person. He takes a bite of a cake and later displays the whole cake. The model may also sometimes confuse spatial details in the video, such as mixing left and right and following a specific camera path.
When is the Sora model available for converting text to video?
While OpenAI has not yet announced the availability of the new artificial intelligence model Sora for creating video from texts, it said that its safety team is now testing the model to ensure potential risks, and a group of creators are now testing it to obtain their comments.
The company said that it will take many steps to ensure safety before making Sora available in its various products, such as ChatGPT Plus And others, while working to develop tools to help detect misleading content. It also said that it will benefit from the current safety methods it has provided for its products, such as DALL E 3 To create images with artificial intelligence by describing them in text.
Examples of videos Sora is able to create with text description:
OpenAI showed several videos of the creation of the Sora text-to-video model, and Sam Altman also showed a group of videos of the text suggestions he got from users on X, which examples included:
Video for text description: “A monkey playing chess in the garden”
Video for text description: “Half duck, half dragon flying through a beautiful sunset with a hamster on his back wearing adventure gear.”
Video for text description: “Homemade gnocchi cooking tutorial hosted by a social media influencer grandma in a Tuscan country kitchen with cinematic lighting.”
In addition to Sora's ability to produce high-quality video clips from text, the new artificial intelligence model is also capable of converting still images into video and extending video clips in any direction, in addition to the ability to edit video clips, change styles and environments, and the ability to merge two video clips together.
