OnpenAI introduced its AI-powered text-to-video generator, Sora, last month. Now, as per reports, the company is planning to launch its latest generative artificial intelligence (AI) model later this year. OpenAI’s Sora can generate 60-second-long videos. In a recent interview, the company’s Chief Technical Officer revealed that Sora can also generate videos of a few minutes in length depending on the complexity of the prompt. She also revealed some details on the imperfection of the model, the data used to train it, and the possible price point of Sora. Here are all the details you need to know.
OpenAI CTO Mira Murati participated in an interview with The Wall Street Journal where when she was asked about the availability of Sora AI video generator, she said, “I'm hoping definitely this year, but could be a few months”.
Further, she was asked if the upcoming US presidential election in November 2024 would affect the company's decision on when to release its AI model, Murati emphasised the importance of dealing with misinformation and harmful bias. She stated that the company would only release the AI model if they were confident about its impact on global elections.
When questioned about the source of data used to train the Sora AI video generator, Murati declined to disclose any specific details. However, she did mention that the AI model was trained on publicly available data, which was licensed to the company.
She confirmed that Sora used content from Shutterstock, with whom they have a partnership, but was unsure whether data from YouTube or Facebook was also used.
According to Murati, Sora is a costly model to operate compared to other OpenAI tools like ChatGPT and DALL-E. Sora is a lot more expensive. The CTO mentioned to the WSJ that the company had not finalised a price point for Sora yet, but they are striving to make it available at a similar cost as DALL-E once it is released.
ALSO READ: AI will be smarter than human by 2025: Elon Musk
