NIST Expands AI Consortium to Speed Healthcare Adoption
The National Institute of Standards and Technology is recruiting organizations to join its renamed AI Consortium, which will focus on building evaluation standards and testing environments for artificial intelligence across healthcare and other sectors. The effort aims to address barriers slowing AI adoption in regulated industries where organizations face unclear governance rules and limited trust in the technology.
NIST announced the expanded consortium in late May, seeking letters of interest from companies and institutions with technical expertise in AI measurement, evaluation and risk management. The institute launched the consortium in 2024 as the Safety Institute Consortium and has now broadened its scope to support "proven, scalable and interoperable uses of artificial intelligence."
Six Task Groups Will Focus on Measurement and Standards
The consortium will organize work around six task groups, each addressing specific gaps in how organizations evaluate and deploy AI:
- AI Testing, Evaluation, Verification and Validation (AI TEVV): Will develop preliminary standards for the private sector.
- Annotation for AI Risks and Validity: Will create a toolkit for assessing AI risks aligned with NIST's ARIA program.
- AI Evaluation and Measurement Methods: Will identify gaps and barriers in AI evaluation science across multiple sectors.
- Bias Effects and Notable Generative AI Limitations (BENGAL): Will explore how large language models can be used effectively for intelligence analysis while addressing misinformation, data leakage and reasoning flaws.
- AI Documentation Cards: Will create standardized templates for documenting AI datasets, models and systems.
- Chemical and Biological Security: Will share insights on emerging AI measurement approaches.
The first review period for applications begins within 60 days of the May 29 notice. Organizations will be selected on an ongoing basis and asked to sign a Cooperative Research and Development Agreement with NIST.
Healthcare Faces Distinct Barriers to AI Adoption
Healthcare has been slower to adopt AI than other sectors, according to the federal government's national AI action plan released last year. The reasons include distrust of the technology, unclear regulatory requirements and a lack of consistent governance standards.
Bill Charnetski, executive vice president of health system solutions and government affairs at PointClickCare, said in a statement: "It's imperative that healthcare adopt policies that allow AI to fulfill its promise while still protecting patient data and preventing misuse."
Health IT vendors have pushed for clearer rules on AI development. The revised NIST consortium addresses this by establishing science-based guidelines and measurement standards that organizations can reference when building and deploying AI systems.
Alignment With National AI Policy
The expanded consortium aligns with the National Artificial Intelligence Initiative Act of 2020 and an executive order signed last year on removing barriers to American AI leadership. The administration has identified medicine and healthcare as sectors where AI breakthroughs are possible.
Craig Burkhardt, deputy NIST director, said: "We are inviting technically capable organizations to join the NIST AI Consortium to address the challenges associated with the development and deployment of AI."
Organizations interested in joining should describe their technical expertise, products, data and models in their letters of interest to NIST.
Learn more about AI for Healthcare and AI for Government.
Your membership also unlocks:
