Spotlight On: It’s Raining AI Devices as Regulatory Storm Reshapes Digital Health

2025 is widely described as the year of AI accountability in healthcare, with regulators shifting from hype to rigorous oversight of AI-enabled medical devices.2

The U.S. FDA has issued new and draft guidance for AI-enabled medical devices emphasizing safety, transparency, real‑world performance, and a total product lifecycle (TPLC) approach, requiring detailed documentation for every AI submission.2 1

The FDA’s public AI‑Enabled Medical Device List now includes hundreds of authorized products and is used as a transparency tool to track AI in clinical care.7 1

Globally, there are now roughly 950+ FDA‑cleared AI/ML medical devices as of mid‑2024, illustrating the rapid growth that is driving more stringent oversight.1

In November 2025, FDA’s Digital Health Advisory Committee focused on generative AI‑enabled digital mental health devices, stressing physician/human oversight and prioritizing enforcement where patient harm risk is higher.3

The EU AI Act now classifies most medical AI as high‑risk, adding obligations for logging, robustness, human oversight, and risk management on top of MDR/IVDR device rules, effectively tightening the regulatory environment for digital health AI.1

The WHO and international bodies such as IMDRF and ISO have published principles and standards pushing for trustworthy AI in health—focusing on transparency, lifecycle management, bias mitigation, and safety—which many national regulators are using as a foundation.1

Professional and industry groups (e.g., AMA, AHA) are pressing for coordinated, whole‑government AI strategies that balance innovation with safety, clarify AI liability, and ensure clinicians are central to AI governance and post‑market surveillance.4 6

New HHS and federal strategies position AI as central to health operations but pair this with strict risk management and compliance checks for high‑impact AI systems affecting health outcomes and sensitive data.5

Regulators are increasingly distinguishing between low‑risk wellness apps (where FDA may exercise enforcement discretion) and higher‑risk clinical tools, focusing oversight on use cases with significant potential for patient harm, particularly in mental health and diagnostic support.3