Daniel Hallenbeck, vice president of strategy and strategic partnerships at Acentra Health
Photo: Daniel Hallenbeck
As AI capabilities advance rapidly, states are under increasing pressure to do more with less, modernize legacy systems and improve operations – often without Medicaid-specific guidance on how to use AI safely and effectively.
The Safe AI in Medicaid Alliance is a public-private collaboration focused on helping Medicaid programs confidently and responsibly adopt artificial intelligence.
SAMA builds on existing federal guidance, particularly the NIST AI Risk Management Framework, which provides a strong but intentionally general foundation. SAMA's self-appointed role is to translate that guidance into a Medicaid context by developing practical resources – such as a Medicaid-specific AI risk overlay, a model state policy and shared use-case frameworks – that states can actually use.
A shot in the arm
At its core, SAMA exists to give states confidence, explained Daniel Hallenbeck, vice president of strategy and strategic partnerships at Acentra Health, a government program health IT vendor. Hallenbeck oversees the company's role in the Safe AI in Medicaid Alliance. He previously was the chief technology officer at New York State Medicaid.
"By creating a neutral space where Medicaid leaders and industry partners can work together, SAMA helps establish a shared understanding of AI benefits, risks and responsibilities," he said. "The goal is not to slow adoption, but to help states move faster – with clarity, consistency and trust so they can realize artificial intelligence's benefits in support of Medicaid's mission.
"In Medicaid, AI's greatest near-term value is as a support tool for modernization efforts already underway," he continued. "When applied thoughtfully, AI can help agencies manage scale and complexity, for example, by assisting with document processing, workflow triage, data validation and pattern detection while keeping human decision making firmly in place."
AI also can strengthen auditability when it is designed with transparency in mind, he added.
Easier to review, not harder
"Systems that incorporate explainability, logging and clear decision pathways can make processes more consistent and easier to review, rather than harder," he said. "In this way, AI can enhance visibility into how work is performed across large and complex programs.
"The key is alignment," he noted. "AI should be deployed in service of program objectives and operational realities. When paired with clear frameworks like the NIST AI Risk Management Framework, AI becomes a practical modernization accelerator, not a risk to accountability."
Hospitals and health systems can navigate the implementation of safe, transparent AI through shared guardrails, defined risk tiers and human-in-the-loop approaches, Hallenbeck said.
"While SAMA is focused on Medicaid programs, many of the principles are applicable across healthcare settings," he noted. "Frameworks like the NIST AI Risk Management Framework provide a common foundation for thinking about AI use, risk and responsibility regardless of organization type.
"From that foundation, shared guardrails – such as distinguishing lower-risk uses from higher-impact ones and ensuring appropriate human involvement – help organizations adopt AI in a consistent and transparent way," he continued. "While the terminology 'human-in-the-loop' has become widely accepted, my team and I believe the industry's focus must be on a 'human-in-the-lead' approach."
Not replacing human judgment
This shift enables AI to support operations and decision-making without replacing human judgment, Hallenbeck said.
"What matters most is having a clear, shared understanding of how artificial intelligence is being used and why," he stated. "Using established frameworks and common language helps healthcare organizations move forward responsibly, even as specific applications and environments differ.
"One clear insight from SAMA is that AI is already being used across Medicaid programs, often cautiously and without much visibility," he continued. "States are exploring tools to support operations, analytics and planning – but many are hesitant to scale without clearer, Medicaid-specific guidance."
Multiple challenges
Capacity is another recurring theme.
"States consistently point to limited staff training, uncertainty around vendor claims and questions about how to apply federal guidance in a Medicaid context," Hallenbeck said. "These challenges are less about the technology itself and more about having the right frameworks, skills and shared understanding in place.
"At the same time, there is strong momentum and interest," he continued. "States want practical examples, common language and reusable tools they can adapt rather than build from scratch. That demand is what SAMA is designed to address – helping states move from uncertainty to confidence as they incorporate AI into Medicaid programs."
Follow Bill's health IT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
WATCH NOW: In 2026, AI will power change in RTLS
