The Coalition for Health AI (CHAI) has recently launched a new model card registry in collaboration with health system members like Providence, Cleveland Clinic, and Kaiser Permanente. This registry aims to provide a centralized hub for hospitals and developers to access and share information on AI tools, making it easier for healthcare companies to evaluate and choose between validated AI products.
AI technology is increasingly being adopted in the healthcare sector, with providers utilizing algorithms for various tasks such as clinical notetaking, patient communication, and staff scheduling. While the majority of medical organizations are already leveraging AI primarily for administrative purposes, there is a growing interest in utilizing AI for more clinical applications like personalized treatment plans and diagnostic support.
Despite the increasing adoption of AI in healthcare, many organizations struggle to keep up with the rapid changes in the industry and navigate through the numerous AI tools available in the market. Procurement of AI products can be challenging, as health systems need to analyze technical materials and marketing presentations to make informed decisions.
In response to these challenges, CHAI’s model card registry standardizes the presentation of information about AI tools by providing model cards that document the development, capabilities, and limitations of AI models. Model cards serve as a “nutrition label” for AI products, allowing hospitals to access information in one location and AI developers to showcase their products to potential buyers.
The model card template was released for public comment last year, and CHAI is working to ensure that it includes appropriate fields for generative AI products, which are more complex to oversee than predictive models. The group is also exploring ways to transform the model registry into a tool that delivers the right information to users at the right time.
Since its establishment in 2021, CHAI has grown to include 1,300 member organizations, including tech giants like Microsoft, Google, and Amazon. The consortium aims to create a network of quality assurance labs to evaluate AI models and develop best practices for deployment in healthcare settings.
In a landscape where the federal government has taken a hands-off approach to regulating emerging technologies like AI, groups like CHAI are filling the gap by establishing standards and guidelines for responsible AI adoption in healthcare. Other organizations, like the Trustworthy & Responsible AI Network (TRAIN) backed by Microsoft and Nvidia working with clinical AI company Aidoc, are also developing standards to ensure the ethical and safe deployment of AI technologies in healthcare.
Overall, CHAI’s model card registry represents a significant step towards promoting transparency and accountability in the adoption of AI in healthcare, providing healthcare organizations with the tools they need to make informed decisions about AI technology integration.
