AI startup Stability AI has released its latest image-generating model, Stable Diffusion XL 1.0. The model, which contains 3.5 billion parameters, can produce high-resolution images in seconds and offers improved colors, contrast, shadows, and lighting compared to its predecessor. Stable Diffusion XL 1.0 is customizable and can be fine-tuned for concepts and styles. It also excels in text generation, capable of generating legible logos, calligraphy, and fonts. The model supports inpainting, outpainting, and image-to-image prompts, allowing users to create detailed variations of images. However, the open-source version of the model raises ethical concerns as it can be used to generate harmful content. Stability AI has taken steps to mitigate this issue by filtering training data and implementing warnings and blocks for problematic content. The training set of Stable Diffusion XL 1.0 includes artwork from artists who have protested against the use of their work in generative AI models. Stability AI is committed to improving safety measures and respecting artists’ requests. The company is also introducing a fine-tuning feature and partnering with Amazon’s cloud platform, Bedrock, to expand its capabilities. Stability AI faces competition from other AI companies and has been raising funds to support its commercial endeavors.
