Thank you to all who joined us on Wednesday, April 17, 2024 for our monthly all-member meeting, Network News. During this month’s meeting, we shared network updates, member highlights, and discussed important member benefits information. Our spotlight presentation featured an expert panel discussion focused on policies and governance of AI applications in health IT and health care. We discussed important ways that Civitas members can support the appropriate use of AI and assist stakeholders with adoption. Our speakers included:

• Alan Katz, Associate Director of Advocacy and Public Policy, Civitas Networks for Health
• Ahmad Alkasir, Policy Lead, Ellison Institute of Technology
• Shamekka Marty, Patient Advocate, Healthcare Interoperability Expert, Health Union
• Jolie Ritzo, Vice President of Strategy and Network Engagement, Civitas Networks for Health (Moderator)

Summary Takeaways from Panel Discussion on AI Policy and Governance with Health IT and Health Care Applications

Responsibly using AI to improve health data, health IT, and health care will require careful policy and multistakeholder governance.

There are many promising applications of AI being used or considered, a few that came out of our panel discussion include:

• Primary use cases at the patient and provider levels: personalizing treatment and patient engagement, improving response time to medical imaging analysis, and drug discovery
• Secondary use cases: health data research, population health and predictive analysis, assessment of data reliability and data quality, reduction of burdensome processes

Though these applications stand to bring about important efficiencies and necessary improvements, we must be careful not to apply AI before we are ready, and we have to ensure that policies, practices, and governance are in place. There is promise in building from already existing infrastructure from health data collaboratives, HIEs, and emerging HDUs. As noted by Ahmad Alkasir with the Ellison Institute of Technology, “Computers cannot be held accountable. AI models should be in the co-pilot seat rather than the driver’s seat.”

We also explored that as more and more AI models are developed this will likely lead to model performance decay which will require decentralized oversight. Governance will have many complexities and will also need to include many different stakeholders. AI governance models have to include patients. Shamekka Marty with Health Union noted involving patients who have deep experience with health care interactions will ensure that important patient voices stay connected to AI applications; AI cannot speak for humans.

Civitas is paying close attention to AI applications in health IT and health care. We will continue to track state and federal policies and look for opportunities to leverage existing infrastructure to support AI governance. Alan Katz with Civitas highlighted that Civitas is tracking over 17 state AI policies and is preparing to respond to a Senate HELP committee RFI focused on AI in health care.