Skip to main content

Regulatory & Policy

Tracks and explains regulatory developments, policy guidance, and governance standards directly relevant to AI in healthcare. Covers FDA AI/ML SaMD action plans and device authorizations, CMS coverage and reimbursement decisions, WHO ethics and guidance publications, CHAI assurance standards, EU AI Act implications for medical devices, and international regulatory frameworks. Each entry is dated, source-attributed, and contextualizes the significance of the development for clinical practice, procurement, and research. This group serves policy professionals, health system administrators, researchers, and legal/compliance teams who need accurate, current regulatory intelligence. Excludes speculative policy commentary without cited basis. Entries link to primary regulatory source documents.

Issuing Authorities

  • CMS
  • FDA (CDRH, CBER, CDER)
  • FDA (CDRH, CBER, CDER, Office of Combination Products)
  • FDA (with Health Canada and MHRA for guiding principles)
  • NIST, FDA, ONC, HHS OCR
  • ONC (Office of the National Coordinator for Health Information Technology)

Policy Types

  • Guidance document
  • coverage decision
  • draft guidance
  • guidance document

Jurisdictions

  • U.S.
  • U.S. (primary); Canada and UK for joint guiding principles
  • US
  • US / EU

Regulatory & Policy Entries

  • CMS Prior Authorization and AI: How Three Federal Rules Are Reshaping Utilization Management

    Effective Date:

    A structured policy analysis of how CMS-4201-F, CMS-0057-F, and the WISeR Model together constitute a deliberate federal strategy to introduce AI and electronic automation into prior authorization — examining each instrument's regulatory theory, compliance obligations, and the unresolved tensions around AI transparency, financial incentives, and health equity that will define the debate through 2026 and beyond. Written for health policy professionals, payer compliance officers, hospital administrators, and clinicians engaged in utilization management policy.

  • ONC Information Blocking Rule: What It Means for AI Systems, Agentic Workflows, and Healthcare Data Access

    Effective Date:

    A structured policy analysis of how the ONC information blocking framework under the 21st Century Cures Act has evolved into the dominant federal mechanism governing AI data access in healthcare — covering HTI-1 algorithm transparency requirements, the HTI-5 proposed rule's deregulatory paradox, and the active enforcement shift of early 2026 that transforms information blocking from a theoretical compliance risk into an operational AI governance imperative. Written for health policy professionals, compliance officers, EHR developers, and health system AI teams.

  • AI in Medicine: How the U.S. Regulatory Framework Actually Works

    Effective Date:

    A structured overview of how the FDA, ONC, and CMS regulate AI applications in clinical medicine — what gets classified as a medical device, what doesn't, and where the policy gaps remain as of Q2 2026.

  • US / EU

    Artificial Intelligence in Health: The Regulatory Landscape in 2026

    Effective Date:

    A structured overview of the formal regulatory and policy actions shaping artificial intelligence in health as of Q2 2026 — covering FDA guidance updates, CMS reimbursement positions, ONC rulemakings, and the EU AI Act's healthcare provisions, with scope labels and primary source references for each entry.

  • US

    FDA AI/ML SaMD Action Plan (2021): Five Commitments, Key Deliverables, and Implementation Status Through Q2 2026

    Effective Date:

    A structured policy-tracker record tracing the FDA's January 2021 AI/ML Software as a Medical Device Action Plan — its five regulatory commitments, the specific guidance documents each produced from 2021 through early 2025, and what remains unfinished or unaddressed as of Q2 2026. Intended for regulatory professionals, SaMD developers, and policy researchers tracking the FDA's evolving framework for continuously learning algorithms.

  • US

    FDA PCCP Guidance for AI-Enabled Devices: Document Record, Version History, and Current Status (Q2 2026)

    Effective Date:

    A structured regulatory tracker record for FDA docket FDA-2022-D-2628 — the Predetermined Change Control Plan guidance for AI/ML-enabled device software functions — documenting its statutory origin, the material changes from the April 2023 draft to the December 2024 final and August 2025 update, adoption data, and open companion document status as of Q2 2026. Intended for regulatory professionals, policy researchers, and SaMD developers who need a primary-source-anchored reference for this specific guidance document's lifecycle.

  • US

    FDA Predetermined Change Control Plan for AI/ML Medical Devices

    Effective Date:

    A structured reference entry on the FDA's Predetermined Change Control Plan (PCCP) framework — what it permits AI/ML device manufacturers to do, how submissions are structured, what regulatory boundaries apply, and why it matters for post-market device management.

  • CMScoverage decisionU.S.

    CMS AI Reimbursement Policy: Coverage Decisions for AI-Assisted Diagnostics

    Effective Date:

    This article provides a detailed analysis of the 2026 inflection point in Medicare reimbursement for AI-assisted diagnostics, covering new Category I CPT codes, HCPCS G-codes, payment challenges, and operational strategies for radiology administrators, cardiology managers, and health system CFOs.

  • FDA (CDRH, CBER, CDER)draft guidanceU.S.

    FDA's January 2025 Draft Guidance on AI-Enabled Device Software Functions: Nine Documentation Requirements and What They Mean for Submissions

    Effective Date:

    A detailed compliance analysis of FDA docket FDA-2024-D-4488, the January 2025 draft guidance on AI-enabled device software functions, explaining its nine documentation areas, bias and transparency mandates, terminology distinctions, and how it fits with the August 2025 PCCP final guidance and the QMSR effective February 2026—written for regulatory affairs professionals, device manufacturers, and health system procurement teams preparing submissions now.

    Clinical Impact: Manufacturers must address nine documentation areas — including model cards in labeling, demographic subgroup performance testing, and postmarket monitoring plans — in AI-enabled device submissions across all pathways (510(k), De Novo, PMA, HDE, BLA, IDE). Health system procurement teams can now evaluate vendors on model card completeness, subgroup performance data, and postmarket monitoring commitments. Guidance remains in draft as of Q2 2026 but reflects current FDA review expectations.

  • FDA (with Health Canada and MHRA for guiding principles)Guidance documentU.S. (primary); Canada and UK for joint guiding principles

    FDA's Predetermined Change Control Plan Guidance: What It Requires and Why It Matters for Health Systems

    Effective Date:

    The FDA's PCCP framework allows AI-enabled medical devices to update algorithmically after clearance without a new submission — a shift that creates distinct transparency, monitoring, and vendor accountability obligations for health systems, procurement teams, and clinical governance leads who deploy these devices.

    Clinical Impact: PCCP-authorized AI devices may update algorithmically post-clearance without a new FDA submission, meaning clinical staff can encounter changed device behavior and revised labeling without a new regulatory review. Health systems must verify vendor labeling compliance, establish notification workflows for PCCP-covered modifications, include bias and equity accountability requirements in vendor contracts, and confirm alignment with QMSR post-market surveillance obligations (effective February 2, 2026).

  • FDA (CDRH, CBER, CDER, Office of Combination Products)guidance documentU.S.

    Predetermined Change Control Plan (PCCP): What the FDA's August 2025 Final Guidance Actually Requires of Manufacturers

    Effective Date:

    A component-level analysis of the FDA's August 2025 PCCP final guidance for regulatory affairs professionals and AI device manufacturers — covering what each of the three required components must contain, which submission pathways apply, enforcement consequences for deviating from an authorized PCCP, and how the February 2026 QMSR transition raises the compliance bar.

    Clinical Impact: Manufacturers submitting 510(k), De Novo, or PMA applications for adaptive AI/ML devices must include a three-component PCCP meeting granular FDA requirements; Special 510(k) is prohibited as a PCCP vehicle. Deviating from an authorized PCCP after clearance risks adulteration or misbranding findings. The February 2026 QMSR transition adds design-control and record-retention obligations that must be satisfied before any PCCP modification cycle can be executed compliantly.

  • NIST, FDA, ONC, HHS OCRguidance documentU.S.

    NIST AI Risk Management Framework in Healthcare: How GOVERN, MAP, MEASURE, and MANAGE Bridge FDA Voluntary Governance and Algorithmic Accountability

    Effective Date:

    For healthcare compliance officers, clinical AI program leads, and policy professionals, this article maps the NIST AI RMF's four core functions onto the specific accountability obligations created by FDA SaMD lifecycle expectations, ONC HTI-1 transparency requirements, HHS Section 1557 algorithmic bias duties, and emerging state AI laws—explaining why a nominally voluntary framework has become the practical governance standard for defensible clinical AI deployment in 2026.

    Clinical Impact: Health systems deploying clinical AI must operationalize NIST AI RMF's four functions to satisfy mandatory obligations under ONC HTI-1 (enforcement discretion closed February 28, 2026), Section 1557 algorithmic discrimination duties for Medicare/Medicaid participants, and state AI laws in Colorado, Texas, and California. Organizations lacking a formal pre-implementation AI approval process or complete AI inventory—including EHR-embedded AI features—face regulatory exposure across multiple enforcement vectors. The GOVERN, MAP, MEASURE, and MANAGE functions provide the governance scaffold for defensible clinical AI deployment and alignment with the CHAI/Joint Commission RUAIH certification pathway.

  • ONC (Office of the National Coordinator for Health Information Technology)guidance documentU.S.

    TEFCA, FHIR, and Clinical AI: What ONC's Interoperability Infrastructure Means for AI Development and Deployment

    Effective Date:

    This article maps how TEFCA's rapid scale-up to population-scale data exchange, ONC's phased FHIR roadmap, USCDI data element standardization, and HTI-5's proposed deregulatory reset have created a new national health data infrastructure — and explains what health system AI teams and clinical AI developers need to understand about building responsibly on it.

    Clinical Impact: TEFCA's scale-up to ~464 million exchanged records and ONC's phased FHIR roadmap create a national data infrastructure enabling population-scale clinical AI. Health systems and AI developers must align training pipelines to declared USCDI versions, implement SMART on FHIR authorization and CDS Hooks for point-of-care deployment, and navigate TEFCA Common Agreement data use obligations and HIPAA de-identification requirements before FHIR Bulk Data enters any AI training pipeline. HTI-5's proposed removal of 34 legacy certification criteria — if finalized — would create a cleaner FHIR-native baseline for AI product integration.