A new model from OpenAI, called Sora, lets users create detailed video scenes of up to 60 seconds using just text-based prompts. It has been met with a combination of awe and deep concern at its potential applications.
OpenAI, publishing several AI-generated videos and their text prompts, said that it is sharing them “to give the public a sense of what AI capabilities are on the horizon" – adding that it is buiding safeguards, like a “detection classifier that can tell when a video was generated by Sora."
It did not publish a general release date for Sora and said it is still Red Teaming the new engine to reduce risk and optimise output capabilities.
OpenAI said that “Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background [and] understands not only what the user has asked for in the prompt, but also how those things exist in the physical world.”
Data scientists speculated that it may have been trained on extensive synthetic data from tools like graphics game engine Unreal Engine 5.
