Học Tiếng Anh qua Tin Tức

Google has officially introduced Lumiere, an advanced AI model capable of generating high-quality videos from text prompts and static images. Announced at the Google Cloud Next 2025 conference, Lumiere represents a major leap in AI-generated video technology. Unlike earlier AI video tools that produced only short, fragmented clips, Lumiere can create seamless, coherent videos lasting several seconds. Its core innovation lies in a new architecture called Space-Time U-Net, which synthesizes entire sequences in one step, significantly enhancing temporal consistency. Users can input descriptive text or still images, and Lumiere will automatically animate them, offering a range of applications from advertising to entertainment. However, Google has restricted Lumiere's public access for now, citing concerns over potential misuse such as misinformation and deepfakes.