Sora will become publicly available within 2024 – Kommersant

American company OpenAI announced, which will make the realistic AI video generator Sora available to the general public within 2024. Currently, only artists, filmmakers and designers can work with it.

In addition to video generation, the creation of an audio track will be added, which will make the videos even more realistic. The company is also working on the ability to edit videos, since the neural network is not always able to create an accurate image the first time – the function is going through the testing stage.

According to the developers, Sora is trained on publicly available and licensed data, and also uses content from the Shutterstock photo hosting site. Videos will not be allowed to contain images of public figures, and the videos themselves will be watermarked. Sora has high system requirements compared to other AI models, and OpenAI is working to make the neural network available to people at the same price as the DALL-E model (which creates images from text descriptions).

OpenAI presented Sora neural network in February 2024, it is capable of generating high-quality videos up to a minute long using a text description. The neural network can create photorealistic videos and animate static images. According to survey HarrisX, videos created by the Sora neural network turned out to be difficult to distinguish from real ones.

Evgeniy Fedunenko