Sora, OpenAI’s video-generating artificial intelligence (AI) service that the company unveiled last month, is expected to launch later this year.
“Sora will be available to the general public this year,” said Mira Murati, OpenAI’s chief technology officer, the Wall Street Journal reported Wednesday. Mira Murati did not specify the launch date, only saying that it will be available “in a few months”.
Sora, unveiled by OpenAI last month, is a generative AI service that converts text to video. It is currently only accessible to a select group of developers and creators. The company is working to improve the service based on their feedback.
Best sound, editing, and legal
Mira Murati said the company also plans to improve Sora’s sound capabilities. “We will improve the sound quality to add more realism to the videos created by Sora,” she said.
“Currently, Sora videos are not editable, but we are considering adding a feature that will allow users to edit their own videos in the future,” she said.
She also says that Sora “only trains from legally permitted data”, but did not release specific data.
“Sora users will not be able to create images of public figures, such as politicians, singers and movie stars,” she said, adding that the system would also have a watermark allowing distinguish between images of Sora and real images. This initiative is part of OpenAI’s efforts to prevent “deepfakes”, “fake news” and other abuses of generative AI.
Source: “ZDNet Korea”