HHS seeks public input on AI's role in healthcare
Federal health officials want to speed up responsible AI use across U.S. clinical settings. The U.S. Department of Health and Human Services (HHS) has issued a Request for Information (RFI) asking for feedback on how regulation, reimbursement, and research can better support AI adoption that improves outcomes, reduces burden, and lowers costs.
"Artificial intelligence will be a transformative force for good across America," said HHS Deputy Secretary Jim O'Neill. "Our efforts to accelerate AI adoption must reflect the real needs of those developing these tools and delivering care."
What HHS is asking for
The RFI invites comment on how to move AI from pilots into routine clinical use across care settings. HHS is extending its internal "OneHHS" AI strategy to shape future actions across agencies. Leaders responsible for scaling AI can review the AI Learning Path for VPs of Strategy to align organizational planning with these expectations.
- Regulatory fit: How digital health and software rules should evolve to safely accommodate AI-enabled tools without slowing useful innovation.
- Reimbursement: How payment pathways can support practical, cost-efficient AI deployments once clinical value is demonstrated.
- Research and implementation: How federal investments can strengthen implementation science, evaluation methods, and best practices. See resources on AI Research for evaluation and measurement approaches.
- High-acuity use: How to address patient safety in complex environments like critical care, oncology, and emergency medicine.
Interoperability, privacy, and trust
HHS underscored that AI depends on secure data exchange and HIPAA-compliant workflows. Thomas Keane, MD, MBA, assistant secretary for technology policy and national coordinator for health IT, put it plainly: "Data liquidity and trust are essential," adding that AI only helps if patients and clinicians understand and trust how data is used.
- Expect strong emphasis on FHIR-based exchange, API access, and clear data provenance.
- Build privacy-by-design processes and audit trails that stand up to scrutiny.
- Be ready to communicate how models use, store, and protect PHI.
Why this matters for your organization
The RFI is a chance to influence rules that will set the pace for clinical AI. If your team is piloting algorithms for imaging, risk stratification, triage, revenue cycle, or documentation support, your practical experience is exactly what HHS is asking for.
- Reduce friction: Clearer guardrails can shorten time from pilot to production.
- Unlock payment: Defined pathways help viable tools qualify for reimbursement.
- Lower risk: Shared safety practices and evaluation frameworks prevent rework and adverse events.
How to contribute: a practical checklist
- Define use cases: Briefly describe the clinical problem, setting, and target users. Include baselines and desired outcomes (e.g., readmissions, time-to-diagnosis, throughput, staff hours saved).
- Data readiness: Note current EHR integration, FHIR resources in use, API maturity, and gaps in data quality or labeling.
- Safety and governance: Share your model oversight, bias testing, human-in-the-loop steps, monitoring thresholds, and rollback plans.
- Workflow fit: Explain how the tool plugs into clinical operations and handoffs. Avoid added clicks and alert fatigue.
- Evidence plan: Propose pragmatic study designs (A/B, stepped-wedge, pre/post) and metrics you can report in 30/60/90 days.
- Payment and cost: Identify reimbursement codes or pathways you believe apply and where current rules block adoption.
- Security and privacy: Outline HIPAA controls, encryption, access governance, and vendor obligations.
- High-acuity safeguards (if relevant): Contingencies for downtime, model drift, and clinician override in critical settings.
- Workforce impact: Training plan, role changes, and measures to reduce burden, not shift it.
- Submit on time: Follow the RFI instructions and include contact info for follow-up. Link to supporting documents where helpful.
Longer-term signals HHS wants you to weigh in on
- Rising frailty, dementia, and chronic disease demands that stretch care teams.
- Worsening workforce shortages across nursing, primary care, and specialties.
- Needs for ongoing model monitoring, recalibration, and transparent reporting.
- Scalable change management and training so tools actually get used.
Helpful references
For context on privacy and interoperability foundations highlighted by HHS:
Upskilling your team
If you are planning organization-wide AI literacy or role-based training for clinical, IT, or operations staff, you can review curated options here: AI courses by job. Executive and strategy leaders may also find the AI for Executives & Strategy resources useful when aligning governance and implementation plans.
Bottom line: HHS is listening. Share what helps you deploy AI safely, what slows you down, and what evidence you can produce quickly. The sooner the field converges on clear, workable patterns, the sooner patients and clinicians benefit.
Your membership also unlocks:
