A China-based startup, Sand AI, has launched an brazenly licensed video-generating AI mannequin that’s garnered reward from entrepreneurs like Microsoft Research Asia founding director Kai-Fu Lee. But Sand AI seems to be censoring photos which may elevate the ire of Chinese regulators from the hosted model of the mannequin, in response to TechCrunch’s testing.
Earlier this week, Sand AI introduced Magi-1, a mannequin that generates movies by “autoregressively” predicting sequences of frames. The firm claims the mannequin can generate high-quality, controllable footage that captures physics extra precisely than rival open fashions.
👀 You gained’t consider that is AI-generated
🧠 You gained’t consider it’s open-source
🎬 You gained’t consider it’s FREEMagi-1 video mannequin simply humiliated business video instruments
Details and examples under: 👇 pic.twitter.com/zlXRecWeqH
— Farhan (@mhdfaran) April 21, 2025
Magi-1 is just too impractical to run on most client {hardware}. It’s 24 billion parameters in measurement, and requires between 4 and eight Nvidia H100 GPUs to run. (Parameters are the inner variables fashions use to make predictions.) For many customers — this reporter included — Sand AI’s platform is the one place they’ll check drive Magi-1.
The platform wants a “immediate” picture to kick off video era. Not all prompts are permissible, TechCrunch rapidly found. Sand AI blocks picture uploads of Xi Jinping, Tiananmen Square and Tank Man, the Taiwanese flag, and insignias supporting Hong Kong liberation. Filtering seems to be taking place on the picture degree; renaming picture recordsdata didn’t skirt the blocking.
Sand AI isn’t the one Chinese startup stopping uploads of politically delicate photos to its video era instrument. Hailuo AI, Shanghai-based MiniMax’s generative media platform, blocks images of Xi Jinping as properly. But Sand AI’s filtering seems to be significantly aggressive; Hauiluo permits photos of Tiananmen Square.
As Wired defined in a bit from January, fashions in China are required to observe stringent data controls. A 2023 regulation forbids fashions from producing content material that “damages the unity of the nation and social concord” — that’s, counters the federal government’s historic and political narratives. To comply, Chinese startups typically censor their fashions, both via prompt-level filters or fine-tuning.
Interestingly, whereas Chinese fashions have a tendency to dam political speech, they typically have fewer filters than their American counterparts for pornographic content material. 404 just lately reported that various video turbines launched by Chinese corporations lack fundamental guardrails that forestall folks from producing nonconsensual nudity.
