Sora, OpenAI’s tool, will soon convert texts into realistic videos. It is accessible to everyone, enriches videos with audio and offers editing options. Despite concerns about data and disinformation, Sora promises to be a major innovation.
Artificial intelligence opens a new chapter. OpenAI, creator of ChatGPT and pioneer in the field, recently unveiled Sora, a text-based video generator that promises to revolutionize the way we express our visual ideas. This technology, which until now was reserved for a handful of people, will become so soon available to the general public.
OpenAI finally provides news about the release of Sora, the AI for videos
Mira Murati, Technical Director of OpenAI, just announced: Sora could be available in a few months. Thanks to these major advances, anyone, regardless of their technical skill level, can create videos.
Imagine: from a simple textual descriptionSora brings hyper-realistic scenes to life, as if your imagination is taking shape before your eyes. And the use of Sora is not limited to artistic creations. Think of scientific simulations, product prototypes, film storyboards, educational materials and much more. The possibilities seem endless
To our colleagues from The Wall Street JournalMira Murati explains that Sora will also be enriched by it audio functions for an even deeper immersion. The company says it will even be possible to edit videos on the fly (like re-running a prompt on ChatGPT to adjust text, for example).
OpenAI aims to make it accessible at affordable prices, “ comparable to that of DALL-E », his famous image generation model. To use DALL-E, you have to pay for ChatGPT Plus, a subscription that costs about 20 euros.
A launch that carries the risk of talking
When Mira Murati was asked about the data used to train Sora, she remained vague, explaining that it was publicly available or licensed data. His only confirmation concerns using content from Shutterstockpartner of OpenAI.
As the US presidential elections approached, Murati indicated that Sora was leaving probably won’t be able to create images of public figures, which adopted DALL-E’s policy. The videos also have a watermark to distinguish them from real images. A solution that is far from infallible.
- Sora is the new OpenAI tool that can create videos from text
- The release date is scheduled within a few months, the company just revealed
- It will be able to make sound and it will be possible to edit videos.