A new AI model has demonstrated the ability to generate realistic video from text descriptions, marking a significant leap in generative AI capabilities. The system, developed by researchers at OpenAI, can create short, high-definition clips based on prompts, showcasing improved temporal coherence and detail over previous methods. While the technology is currently limited to internal testing due to concerns about potential misuse, it highlights rapid progress in multimodal AI. Experts note the advancement brings both creative possibilities and heightened questions about deepfakes and digital authenticity. The full details of the model’s architecture and training data are outlined in the research paper. Read the full article at https://technologyreview.com/2024/05/20/ai-video-generation-sora-advances.