North American artificial intelligence research firm OpenAI has announced its video generation model “Sora.”
According to the official OpenAI website, this model can be used to create a one-minute video from text. A video can contain complex scenes, such as multiple characters, certain types of action, subtle themes, and background details.
The company said it developed the model to teach artificial intelligence to understand and simulate the moving physical world and train it to help people solve problems that require interaction with the real world.
CNN quoted Hayden, a senior analyst at market research firm ABI Research, as saying that the Sora movie could have a major impact on the digital entertainment market because the new personalized content will be disseminated across multiple channels.
In the introduction to “Sora,” OpenAI also pointed out that the current model has weaknesses. The physical performance of complex scenes may be difficult to accurately simulate and may not be able to understand specific instances of cause and effect. For example, you could potentially show a video where after a person takes a bite of a cookie, there are no bite marks on the cookie.
At the same time, the model may also confuse spatial details such as left and right.
OpenAI said it plans to work with a team of experts to test the latest models and stay alert for several areas, including misinformation, hate content and bias.
“Coffee trailblazer. Social media ninja. Unapologetic web guru. Friendly music fan. Alcohol fanatic.”