OpenAI announced public availability later this year Sora, its AI capable of creating realistic videos starting from text prompts or images. He confirms it Mira MuratiChief Technology Officer di OpenAIwho, without providing a precise date, explained that the new AI tool is about to arrive.
OpenAI announces availability of Sora later this year
In an interview with the Wall Street Journal (via Hardware Upgrade), Murati assured that the launch of Sora it will happen before the end of the yearalthough it may take a few months for general availability.
Presented for the first time last February, OpenAI Sora has demonstrated that it can start from simple textual descriptions to produce videos that, at least at first glance, seem truly realistic. An expert eye could easily recognize some problems — from bad reflexes to human hands (which the AI proves to struggle to understand. But despite some problems, the first results have impressed.
OpenAI plans to expand the capabilities of Sora integrating audio functionality, with the aim of making the generated scenes even more realistic. Furthermore, Murati will allow you to edit the content of the videos produced by Sora. In order to remedy the mistakes that AI continues to make.
Murati confirmed that at the time of the public launch, Sora will not be able to generate images of public figures, following the same policies adopted for GIVE HER. Additionally, videos generated by Sora will be marked with a watermark to distinguish them from real content. A way to limit the use of Sora to create deepfakes and potential fake news.
For the moment there is still no date: we will keep you updated.
Stay updated by following us on Google News!
Follow!
Source
The Wall Street Journal