OpenAI unveils Sora; Artificial intelligence model for converting text to video

Last night the company OpenAI unveiled a new model for video production called Sora, which can create incredibly realistic videos using text commands. This company has announced that the model in question is able to deliver videos up to a maximum of 1 minute.

According to OpenAI's announcement in a post dedicated to introducing the model, Sora is capable of creating complex scenes with multiple characters, a variety of specific movements, and detailed subject and background details. In addition to producing video based on text, Sora model is able to create video based on a still image or fill the missing frames of an existing video and even increase its length. Although with a little attention we can notice that the videos are artificial, but it offers amazing videos compared to similar models.

At first, these artificial intelligence models were text-to-image conversions that made a lot of noise. But in the last year, we've seen an amazing amount of progress in text-to-video conversion. Companies like Runway and Pika have shown impressive text-to-video models, and Google's Lumiere is one of OpenAI's main competitors. Similar to Sora, Lumiere offers users text-to-video tools and also allows them to create a video from a still image.

Currently, access to the Sora model is only possible for people in a group known as the “Red Team” who assess the damage and potential risks of this model. OpenAI has also given some artists, designers and filmmakers access to the model for feedback. The company emphasizes that the current model may not correctly simulate the physics of a complex scene and may not correctly interpret some cause and effect cases.

Source: The Verge