OpenAI is set to launch Sora, an AI creating hyperrealistic videos from text

From Merca2.0:

OpenAI introduces Sora, an AI transforming text into hyperrealistic videos, set to launch in a few months. The model promises immersive experiences with advanced features like video editing and audio addition.
OpenAI takes an ethical stance, using a mix of free and licensed data to train Sora, addressing concerns of AI model training exclusively on open-access data.
To counter misinformation, Sora will be limited in generating representations of famous characters, safeguarding against potential false content creation. Sora is OpenAI’s first multimodal language model capable of generating lifelike videos from textual input.
Sora utilizes a diffusion model to create videos from static noise, offering realistic and fluid results in minutes. Despite its advancements, Sora is still improving in areas like cause and effect understanding and spatial awareness.
OpenAI is on the brink of revolutionizing content creation with Sora, opening doors to innovative audiovisual creations. The technology community eagerly anticipates the endless possibilities and impact this advanced AI will have across various industries.



Read more at Merca2.0: When will Sora OpenAI be available? What we already know about its release date