A new AI model, developed by researchers at a leading university, demonstrates the ability to generate realistic video from simple text descriptions. The system, named ‘VidGen’, uses a novel diffusion-based architecture to create short, coherent video clips that match user prompts. While the technology shows significant progress in video synthesis, the researchers acknowledge current limitations in resolution and clip duration. They emphasize the model’s potential applications in film pre-visualization and creative tools, while also highlighting the need for robust safeguards against misuse. The full research paper is available for peer review. Read the full article at https://technologyreview.com/2024/03/15/vidgen-ai-video-model.