Health and Human Services (HHS) AI Use Case

The FY25 AI use case inventory from the U.S. Department of Health and Human Services offers a clear view into how one of the federal government’s largest and most complex agencies is approaching artificial intelligence. What emerges is a picture of growing discipline in governance and documentation, paired with a persistent challenge around scale and operational integration.

Compared to prior years, the FY25 inventory reflects meaningful progress in how AI systems are described and framed. Use cases are more clearly articulated, with stronger explanations of system purpose and intended outcomes. There is also more consistent use of language around risk, oversight, and human involvement. Across HHS components, teams appear to be aligning around a shared understanding of how AI should support decision-making rather than replace it.

That consistency matters and suggests that department-wide guidance on responsible AI is beginning to take hold. When operating divisions describe safeguards, human review processes, and system limitations in similar ways, it points to a maturing governance framework. In a decentralized agency like HHS, that level of alignment is difficult to achieve.

At the same time, the inventory reveals a familiar friction point. While governance language is becoming more standardized, many systems remain limited in scope. Use cases are often focused on internal analytics, research support, or narrowly defined decision support. Others appear to be in pilot or early-stage deployment, with limited evidence of expansion into broader program operations.

This gap between governance maturity and operational scale is where the real challenge lies. Defining responsible AI is only one step in the process. The next is embedding it into day-to-day program execution, which requires sustained ownership, integration with existing workflows, and confidence from program leaders that these tools can deliver reliable outcomes.

The inventory also raises questions about continuity beyond initial development. Few use cases demonstrate long-term operational ownership or lifecycle management. Without that, even well-designed systems risk remaining stuck in pilot phases rather than evolving into mission-critical capabilities.

For HHS, the path forward is not about rethinking governance. The department has made progress in establishing a common foundation for responsible AI. The focus now must shift toward execution by scaling viable use cases, investing in supporting infrastructure, and ensuring accountability extends beyond development into sustained operations.

The FY25 inventory ultimately suggests that HHS understands what responsible AI should look like. The harder challenge ahead is translating that clarity into systems that are fully embedded, operationally owned, and capable of supporting real programs at scale.

Talk to one of HumanTouch's SMEs on AI governance at info@humantouchllc.com for more information.