Creators of ChatGPT launch neural network that generates realistic videos based on text

The American technology company OpenAI, the creator of the intelligent chatbot ChatGPT, has announced the launch of a new artificial intelligence (AI) tool called Sora. This information was provided on the official website of the company.

According to the description, Sora "can create realistic and creative scenes based on textual instructions" with a duration of up to one minute.

The company emphasized that the AI model represents how described objects exist in the real world and can accurately interpret props, creating "convincing characters expressing vivid emotions."

The neural network is also capable of generating videos based on still images, creating missing frames in finished video clips, or complementing them. OpenAI warned that Sora may make mistakes in modeling complex scenes.

Currently, the AI tool is being tested with so-called "red teaming," which evaluates the model for potential harm and risks.

"We will engage politicians, educators, and artists worldwide to understand their concerns and identify positive use cases for this new technology," stated OpenAI.

Here are some examples of videos.

via GIPHY

via GIPHY

via GIPHY

via GIPHY

via GIPHY