OpenAI plans to unveil its latest generative artificial intelligence (AI) model Sora later this year, according to a report. Sora, unveiled last month, is an AI-powered text-to-video generator that the company claims can generate up to 60-second videos, far longer than any industry competitor. In a recent interview, the AI ​​firm’s CTO also revealed that Sora can generate videos within minutes depending on the complexity of the prompt. The interview also covered topics such as the imperfection of the model, the data used to train it, and the possible cost of Sora.

In an an interview with The Wall Street Journal, OpenAI’s CTO Mira Murati, when asked when the AI ​​video generator would be available, said, “I hope definitely this year, but it could be a few months.” When asked if the US presidential election, which are scheduled for November 2024, will play a role in the company’s considerations of when to release the AI ​​model, Muratti said tackling misinformation and harmful bias is important to the company, adding: “We won’t release anything we don’t feel confident about how it might affect global elections.”

Murati was also asked about the source of the data that Sora was trained on. The matter is also of interest to Italy’s data protection watchdog, which is investigating possible data breaches by OpenAI. Regulators gave the AI ​​firm 20 days to provide information about the AI ​​video generator.

In the interview, Muratti also didn’t seem to divulge any details and simply stated that the AI ​​model was trained on data available in the public domain and licensed to the company. She also said she wasn’t sure if YouTube or Facebook data was used, but confirmed to the publication that Sora used content from Shutterstock, with whom they have a partnership.

Sora is also a very expensive model to run compared to other OpenAI tools like ChatGPT and DALL-E, Muratti explained. “ChatGPT and DALL-E are optimized for the public to use, while Sora is really a research output. It’s much, much more expensive,” the CTO told the WSJ. While the company hasn’t finalized a price point for the AI ​​video model, Muratti said the company is trying to make the model available once it’s released at a similar price point to the DALL-E. Currently, the base price for the DALL-E 3 model is $0.040 (roughly Rs. 3) for an image with a resolution of 1024 x 1024 pixels.

In addition, OpenAI’s CTO also emphasized that the company is working to provide Sora’s audio capabilities, optimize generation time and its cost-effectiveness, and make further improvements to minimize imperfections.

Affiliate links may be automatically generated – see our ethics statement for details.