This AI can create 'videos of the future'

Researchers from the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a deep learning system that analyses a static photo, predicts what will happen next, and produces short video clip of this anticipated future – or what they call "Creating Videos of the Future." The system was trained on 2 million unlabeled videos, or about two years of footage. When feeding it with a photo, the system generates a video predicting the next couple seconds in the photo's scene. These algorithm-generated videos were deemed realistic by 20% of human subjects, based on 13,000 opinions from 150 users. This technology and similar technology will have a range of applications in the future.