New alliance hopes to steward responsible AI in Medicaid services
MarioGuti via Getty Images
More than 30 states have joined a public-private initiative that supports collaboration among Medicaid and tech industry leaders to develop responsible AI frameworks and guidelines.
A new public-private initiative aims to offer state leaders guidance and information-sharing opportunities in their efforts to implement artificial intelligence into critical services like Medicaid.
The Safe AI in Medicaid Alliance, launched earlier this month, brings together state Medicaid officials and industry leaders to develop responsible approaches to AI’s use in the health care program, said Verlon Johnson, chief strategy officer at Acentra Health.
Acentra Health’s Healthcare Advisory Board helped create the alliance, in partnership with the Washington State Health Care Authority, the Arizona Health Care Cost Containment System, Amazon Web Services, McKinsey & Co. and Arcadia. Currently, the alliance includes 32 state members and nearly 100 industry leaders, Johnson said.
Artificial intelligence “has this enormous potential to make Medicaid work better, faster and fairer, and we've seen that other industries — banking and others — have really embraced the concept of AI,” she said. “But in a program like Medicaid that is critical, innovation without guardrails is risky, because we're talking about … a vulnerable population.”
More than 70 million Americans are enrolled in Medicaid, according to federal enrollment data from May. That population includes low-income individuals, people with disabilities or new moms, Johnson said. And while AI can help make Medicaid processes more efficient in areas like identifying fraud, the tech can also pose privacy or security risks when not deployed with caution.
“Medicaid agencies are exploring AI to improve administrative efficiency, streamline eligibility and enrollment processes, enhance fraud detection and support clinical decision-making,” a spokesperson for AHCCCS, Arizona’s Medicaid agency, said in an email to Route Fifty.
AHCCCS, for instance, is working on the modernization of its Medicaid Enterprise System to better serve members and providers. SAMA could help the agency’s efforts to evaluate AI tools that help automate repetitive tasks, improve data accuracy and support member engagement, while maintaining compliance with privacy and equity standards, they said.
Through the alliance, for example, members will have opportunities to collaborate to determine how to deploy AI systems that still prioritize human decisions, create guardrails for protecting beneficiaries’ information, develop tools and shared documents to classify appropriate AI use cases and create roadmaps to guide states’ adoption of AI, Johnson explained.
In early conversations with state members, many Medicaid officials were eager to explore AI’s potential to enhance the health care program but reported facing difficulty getting legislators to trust the use of the technology, Johnson said. SAMA can help address such hesitancy by encouraging collaboration among partners to determine safe and responsible frameworks for AI’s adoption and use.
The alliance “is all about shared learning” to support intrastate peer-to-peer collaboration and transparency among vendors that states are working with, Johnson said.
The launch of the alliance comes amid an absence of federal frameworks for AI and the rise of various regulatory efforts from states, which some experts have said could create a patchwork of policies that complicate nationwide AI innovation.
“SAMA provides a collaborative space for states to share best practices, define risk tiers, and co-develop evaluation frameworks tailored to Medicaid’s unique needs,” the AHCCCS spokesperson said. “By working together, states can avoid duplicative efforts and ensure that AI tools are deployed safely, effectively, and in ways that improve outcomes without creating new risks.”