What may two dogs making radio programs (节目) look like? How about a basketball match with different sea animals as players?
With Sora, you may find out quickly. The latest AI model can make videos from some key words in seconds. It was made by OpenAI, the company (公司) behind the popular chatbot (聊天机器人) ChatGPT.
Most people cannot use Sora now. OpenAI is asking some experts to test its possible risks (风险). The model is also open to a small group of artists. The company hopes to make Sora better.
Before Sora, the leading text-to-video AI model was made by Runway. The videos it made were short and frightening. So most people were surprised by the high quality of Sora's videos.
Sora model has some problems now because it may not understand cause and effect(因果关系). "For example, a person may eat a biscuit , but in the video the biscuit may not have a bite mark(咬痕)," OpenAI says on its website. What's more, Sora cannot tell left or right. At the same time, some people have other worries. They are afraid that some people may not use AI in a correct way. Unreal AI-made information is the biggest risk the world faces. To solve this problem, OpenAI says that it is building tools to help users to find out videos made by Sora.