Google has announced a new AI system called ‘Whisk’ that can generate images from text prompts, similar to tools like DALL-E and Stable Diffusion. Whisk uses a novel approach called ‘Diffusion with Semantic Guidance’ that allows for more control over the generated images. Users can provide detailed text descriptions, and Whisk will create corresponding visuals. The system is designed to handle complex prompts and generate high-quality, coherent images. Google claims Whisk outperforms existing text-to-image models in terms of image quality and adherence to the prompt. The company plans to release Whisk to the public in early 2025, with potential applications in fields like art, design, and media production. However, concerns have been raised about the potential misuse of such powerful AI image generation tools.
Source: https://www.cnn.com/2024/12/17/business/google-ai-whisk-image-prompts/index.html