FDA DHCoE: De-Risking AI Investments in Digital Health

The regulatory field for artificial intelligence in healthcare is experiencing an unprecedented acceleration, driven by both technological innovation and legislative mandates. At the heart of this evolution is the Food and Drug Administration’s Digital Health Center of Excellence (DHCoE), a specialized branch tasked with working through the complexities of software as a medical device (SaMD) and the burgeoning field of generative AI. For regulatory affairs directors and digital health investors, understanding the DHCoE’s shifting operational focus and upcoming priorities is not merely advantageous, but critical for strategic planning and de-risking investments.

The Evolving Mandate: Congressional Directives and Generative AI

The DHCoE, since its inception, has been instrumental in shaping the regulatory pathway for digital health products. However, recent congressional mandates, particularly those stemming from the Food and Drug Administration Omnibus Reform Act (FDORA), are significantly broadening its scope and intensifying its workload. These legislative directives demand greater clarity and predictability for innovative digital health solutions, particularly those using advanced machine learning and generative AI. The surge in generative AI submissions presents a unique challenge, as these models often exhibit characteristics like emergent behavior and continuous learning, which can complicate traditional regulatory frameworks. The DHCoE is now tasked with developing nuanced approaches that balance rapid innovation with patient safety and efficacy. This includes grappling with issues like algorithmic drift, where models degrade over time as real-world data distributions shift, necessitating strong post-market surveillance strategies.

Strategic Priorities and Budgetary Allocations

The DHCoE’s strategic plan reflects a clear pivot towards proactive guidance development and stakeholder engagement to address these new realities. While specific budget allocations for the DHCoE are embedded within broader FDA appropriations, an observable trend indicates increased resourcing dedicated to digital health initiatives. This investment is important for the DHCoE to expand its technical expertise and operational capacity to evaluate the growing volume and complexity of AI-driven medical devices. Key priorities include developing frameworks for the responsible development and deployment of generative AI in clinical settings, promoting good machine learning practice (GMLP) principles, and fostering the use of real-world evidence (RWE) for regulatory submissions and post-market monitoring. For investors, understanding these budgetary signals can indicate areas of future regulatory focus and potential market growth.

Upcoming Guidance and Policy Deadlines

Regulatory affairs directors should be aware that several key FDA guidance documents from the DHCoE have recently been issued, while others are actively being developed. These guidances address critical areas such as:

• Predetermined Change Control Plans (PCCPs): The FDA issued final guidance on “Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions” in August 2025. This guidance refines the requirements for PCCPs, enabling AI/ML-driven SaMDs to make predefined modifications without requiring new premarket submissions for every iteration. This is an important area for AI-native companies whose products are designed for continuous improvement.
• Generative AI in Medical Devices: The FDA’s Digital Health Center of Excellence issued a discussion paper titled “Considerations for the Regulation of Generative AI-Enabled Medical Devices” on August 18, 2026, seeking stakeholder input. This indicates ongoing development of foundational guidance on unique considerations for generative AI, including data provenance, bias mitigation, transparency, and validation methodologies. This will be essential for companies developing AI that generates new data, images, or diagnostic interpretations.
• Clinical Decision Support (CDS) vs. Diagnostic AI: Revised FDA guidance for clinical decision support (CDS) was published on January 6, 2026, superseding the 2022 version. This updated guidance clarifies the distinction between CDS tools, which offer recommendations and may be unregulated, and diagnostic AI, which makes independent determinations and is regulated as a medical device. This distinction remains a critical determinant of regulatory pathway and compliance burden.
• Cybersecurity for AI/ML Devices: Final cybersecurity guidance was issued in June 2025, strengthening requirements across all premarket pathways and extending secure-by-design obligations to training data and model artifacts. Given the increased attack surface presented by complex AI systems, enhanced guidance on cybersecurity best practices, including threat modeling and incident response for AI/ML SaMDs, is highly probable.

These documents provide much-needed clarity for companies working through the 510(k) clearance and De Novo classification pathways, especially for novel AI functions that lack established predicates.

Actionable Timelines for Compliance Teams

For compliance teams preparing new submissions, the DHCoE’s evolving agenda shows the need for proactive engagement and a forward-thinking approach.

“The era of reactive compliance is over. Companies must anticipate regulatory shifts and integrate GMLP principles and strong data governance from product inception. The FDA is moving towards a lifecycle approach, and your regulatory strategy needs to reflect that.”

Specifically, compliance teams should:

• Monitor FDA’s Unified Agenda: This public document provides insights into planned regulatory actions and guidance development. Regular review will help identify upcoming policy shifts. FDA Unified Agenda
• Engage with DHCoE Public Workshops and Comment Periods: These forums offer invaluable opportunities to understand the agency’s thinking and provide industry input, potentially influencing the final shape of guidance documents.
• Prioritize Data Moat and Quality Management Systems: For digital health investors, a company’s ability to demonstrate a strong data moat and adherence to rigorous quality management systems (QMS), such as ISO 13485, will be increasingly scrutinized. These are not merely operational necessities but significant de-risking factors for regulatory approval and market adoption.
• Prepare for Enhanced Post-Market Surveillance: As AI models evolve post-deployment, the DHCoE will increasingly focus on real-world evidence collection and monitoring for algorithmic drift. Companies must establish strong systems for continuous performance assessment and planned updates.

The DHCoE’s focus on structured data, transparent algorithms, and continuous monitoring directly impacts the investment case for healthcare AI. Companies demonstrating a clear understanding of these forthcoming requirements, and integrating them into their product development and regulatory strategies, will be best positioned for success.

Methodology and Source Note

The insights presented in this news brief are compiled from an analysis of the FDA’s publicly available regulatory agenda, strategic plan documents from the Digital Health Center of Excellence, and recent legislative mandates from the US Congress. Our editorial team continuously tracks these developments to provide timely and actionable intelligence for our target audience. DHCoE Strategic Plan FDORA legislative text