Overview of the Initiative
The Coalition for Health AI (CHAI) is launching a certification program for independent assurance labs that will evaluate AI algorithms used in healthcare. This initiative aims to create a standardized output, referred to as “model cards,” which will function similarly to ingredient labels, providing essential information about AI tools. The certification process is expected to be finalized by April 2025, following a period of public and member feedback. This move responds to the growing reliance on AI in health systems and the need for transparency and safety in these technologies.
Key Details
- CHAI plans to establish a network of up to 30 assurance labs to vet AI algorithms.
- The certification framework is developed in collaboration with the ANSI National Accreditation Board and emerging quality assurance labs, adhering to ISO 17025 standards.
- Model cards will include critical information such as developer identity, intended use, data types, performance metrics, and known risks.
- The program integrates data quality standards from the FDA and aligns with the National Academy of Medicine’s AI Code of Conduct.
Importance of the Initiative
This certification program is crucial for ensuring that AI tools are safe, effective, and free from bias. With common definitions and standards, health systems can engage in informed discussions with technology vendors. This transparency is vital for building trust and ensuring that AI applications do not harm patients or perpetuate discrimination. By fostering an environment of accountability, CHAI’s initiative aims to improve the overall quality of AI in healthcare, ultimately enhancing patient care and outcomes.
