OpenAI has officially announced that its cutting-edge text-to-video generator, Sora, will soon be accessible to the public. In a recent interview with The Wall Street Journal, OpenAI’s Chief Technology Officer, Mira Murati, revealed that It is slated for release “this year,” with a potential launch within the next few months.
Introduced by OpenAI in February, Sora has garnered significant attention for its ability to create realistic scenes based on text prompts. While initially limited to visual artists, designers, and filmmakers, videos generated by this Ai model have already begun appearing on platforms like X.
In addition to the forthcoming public release, OpenAI has ambitious plans to incorporate audio capabilities into Sora, aiming to elevate the level of realism in generated content. Users can also anticipate enhanced editing features to refine their creations, as OpenAI aims to position Sora as a versatile tool for content creation.
Also Read : Midjourney Accuses Stability AI for Data Scraping
When discussing the data utilized to train Sora, Murati emphasized that it primarily originates from publicly available or licensed sources. Notably, Sora leverages content from Shutterstock, with whom OpenAI has established a partnership. Despite the increased resource requirements associated with it, OpenAI aims to maintain its accessibility in terms of cost, similar to their previous model, DALL-E.
According to OpenAI, Sora’s advanced language comprehension enables it to accurately interpret prompts and generate captivating characters that convey vibrant emotions. Furthermore, It can produce multiple shots within a single video, ensuring consistency in characters and visual style.
Addressing concerns surrounding generative AI tools and misinformation, Murati assured that Sora will refrain from generating images of public figures upon release. Additionally, videos generated by Sora will feature watermarks to distinguish them from genuine footage, although questions persist regarding the efficacy of this measure in preventing misuse.
