In addition to developing a new search engine that aims to challenge existing ones, North American company OpenAI, owner of the digital AI chatbot ChatGPT, publicly presented on Thursday its latest invention, Soraa generative AI tool that allows users to create realistic videos or fictional animated scenes lasting up to one minute from written prompts and instructions that act as commands.
According to OpenAI, “Sora has the ability to produce complex scenes with multiple characters, specific types of movement, and precise subject and scene details. The platform can also understand not only what the user asked for in the prompt, but also how those things exist in the physical world.”
Although it's not new, as companies like Google and Meta have unveiled similar technologies in the past, the videos produced by the new tool are of higher quality. However, like competing platforms, it has not yet been launched due to the potential risks the tool could bring.
Speaking to the New York Times, the program's research leaders, Tim Brooks and Bill Peebles, explained that the company will not make Sora available yet because it is still working on improving the system. To achieve this, OpenAI shares the technology with a small group of academics and other external researchers.
In the case of the following video, produced using the new tool, the prompt was “Several giant mammoths approaching as they walk across a snowy meadow, their long woolly fur blowing lightly in the wind as they walk, trees covered in snow and dramatic snow-capped mountains in Distance, mid-afternoon light with fluffy clouds and the sun rising in the distance create a warm glow, and the camera's low view captures a large, furry mammal with a beautiful depth of field.
“Friendly zombie fanatic. Analyst. Coffee buff. Professional music specialist. Communicator.”
