Chinese AI Startup Sand AI Censors Politically Sensitive Images in New Video Model
Sand AI, a startup based in China, has launched its video-generating AI model, Magi-1, which has received acclaim from figures like Kai-Fu Lee. However, testing reveals that the platform blocks uploads of politically sensitive images, including those of Xi Jinping and Tiananmen Square. This censorship aligns with China's stringent regulations on information control, which restrict models from generating content that undermines national narratives. Unlike its U.S. counterparts, Sand AI appears to have fewer filters for explicit content. The model itself requires substantial hardware to operate, making it impractical for average users.
Created at:
