Sora: OpenAI confirms availability of video-generating AI in 2024

Sora, an artificial intelligence (AI) that can create videos based on textual descriptions It should be made available to the public in 2024, According to Mira Murati, OpenAI’s chief technology officer. In an interview with Wall StreetJournalthis Wednesday (13) the manager commented that the vehicle may also gain voice functionality.

The technology, announced by the same ChatGPT developer in February, is currently Available only to a select group including filmmakers, designers and visual artists. In this first phase, early users of the tool contribute to the development of the model ahead of the commercial launch.

Prompt: “A stylish woman walks down a Tokyo street filled with warmly glowing neon and bustling city signs. she is wearing a black leather jacket, a long red dress and black boots, and she is carrying a black bag. She is wearing sunglasses and red lipstick. He walks confidently and comfortably… pic.twitter.com/cjIdgYFaWq

— OpenAI (@OpenAI) February 15, 2024

According to OpenAI’s CTO, this will change and Sora’s rollout won’t take more than “a few months.” But, has not confirmed whether the tool will be released to anyone or whether only certain audiences will be able to use the mechanismAs in the experimental phase.

In other news, Murati commented on some plans to optimize OpenAI’s video renderer. To develop more realistic scenes, the company plans to: Incorporates audio into generative AI-generated content as well as allowing users to edit original materialcorrection of any inaccuracies.

Concerns about misuse of Sora AI

Although few people currently have access to OpenAI’s video AI, some videos created by the tool are being shared on social media. Many of them show a high degree of realism, which enhances the visual appeal. Concerns about the use of technology in creating deepfakes.

As the United States presidential elections approach (they will be held later this year), the potential of AIs could be misused to spread disinformation. Regarding this risk, the CTO said that the OpenAI tool probably won’t be able to create images of well-known people.

Template that also creates realistic text-based videos There will be a watermark to distinguish AI-generated content from real videosAccording to Murati. In this case, the company will need to develop a marker that cannot be easily removed with editing tools, as experts say.