AI video generators can’t understand the laws of physics solely by watching videos, scientists have found.

Coming hot on the heels of chatbots and image generators, AI video generators like Sora and Runway have already been delivering impressive results. But a team of scientists from Bytedance Research, Tsinghua University, and Technion were curious to learn if such models could discover physical laws from visual data without any additional human input.

While in the real world, we understand physics through math, in the world of video generation, an AI model that understands physics should be able to watch a sequence of frames and then predict which ones come next. This should happen both when the images are ones the AI model has seen before and also unfamiliar ones.

To find out whether this understanding exists, the scientists created a 2D simulation using simple shapes and movements and created hundreds of thousands of mini videos for their model to train and be tested on. They found that the models could ‘mimic’ physics but not understand it.

Keep reading at yahoo.com.