OpenAI announced a new AI model, Sora, that creates realistic clips from text

OpenAI, a leading artificial intelligence company, has unveiled its latest development, a generative artificial intelligence model called Sora. This innovative system enables the creation of text-based videos with incredibly realistic visuals.

The developers of Sora note that videos created using the current version of this artificial intelligence model are limited to one minute. Still, some of the generated videos appear so real that they can be confused with real footage shot with a regular camera. It’s certainly possible to spot the AI’s work, but in some cases, especially on a quick glance, you’ll have to try harder.

It’s important to note that Sora’s neural network generates the entire video sequence instead of stitching together individual frames. This allows for smoothness and minimal artifacts in the video. The official description of the project also indicates that the neural network is trained on the physics and behavior of various objects in the real world.

Sora is currently under active development and only some OpenAI employees have access to this neural network. In the next phase, the developers plan to distribute the new tool to artists, filmmakers and designers to get feedback from experienced professionals. Later, such capabilities will probably be available to ordinary users as well.

OpenAI’s strategy is not to leave filmmakers out of work. The company promises that as they refine artificial intelligence, they will consider the challenges and needs of professionals, and how they can expand the use of Sora to address those challenges.