Stability AI Stable Diffusion Launch to Advance Ethical AI
Stability AI announces the public release of Stable Diffusion, after first releasing it to researchers.
Based on the response to the new AI model, Stability AI has implemented a range of improvements to ensure it is released safely and ethically. The company has incorporated data from its beta model tests, in addition to its community, for developers to act on.
Image generation models are powerful, but still need improvements to understand how to better represent what the user wants. Stability AI’s release is the culmination of many hours of work to create a single file that compresses visual information into a few gigabytes.
Stable Diffusion is a significant advancement in AI image generation and is designed to offer both high-quality results and greater accessibility to a wide range of users.
Prioritising AI ethics
In cooperation with technology teams at Hugging Face and CoreWeave, Stability AI has announced the model is being released under a Creative ML OpenRAIL-M to enable both commercial and non-commercial use. This licence is based on the ethical and legal use of the model as the responsibility of the user.
Likewise, the company has developed an AI-based Safety Classifier, which can understand concepts and other factors in generations to remove outputs that may not be desired by the model user. The parameters of this can be adjusted, with Stability AI welcoming input from the community on possible improvements.
“We hope everyone will use this in an ethical, moral and legal manner and contribute both to the community and discourse around it,” the company says in its announcement. “Please carefully read the model card for a full outline of the limitations of this model and we welcome your feedback in making this technology better.”
Stability AI licences are designed to offer flexibility for the user’s individual generative AI (Gen AI) needs by combining its range of open models with self-hosting benefits.
Moving forward, the company states it will release optimised versions of its Stable Diffusion model, in addition to other variants and architectures. These will have improved performance and quality, in addition to optimisations that will work on AMD, Macbook M1/M2 and other chipsets. Currently, Stability AI recommends using Nvidia chips.
“This technology has tremendous potential to transform the way we communicate and we look forward to building a happier, more communicative, and creative future with you all,” Stability AI says.
About Stability AI
Stability.ai is one of the world’s leading open source Gen AI companies. It aims to deliver breakthrough, open-access AI models with minimal resource requirements in imaging, language, code and audio.
Another offering, DreamStudio, is also powered by Stable Diffusion. It is a world-leading AI image generator that offers extensive customisation and free credits and affordable options, with features like adjustable image size, prompt adherence, and diffusion steps. Users can select algorithm versions and use specific seeds for repeatable results. The company’s use of open-source AI enables more specific model fine-tuning.
Aiming to democratise AI and build a global foundation to achieve humanity’s full potential, Stability AI holds a unique vision and approach to AI with its openness. It aims to foster a culture of trust, transparency, innovation and integrity with the technology.
The London-based company is currently one of the leading AI startup companies in Europe. In 2022, it achieved unicorn status when it achieved a US$1bn valuation in 2022.
The company also recently appointed Prem Akkaraju as its new CEO.
******
Make sure you check out the latest edition of AI Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
******
AI Magazine is a BizClik brand