FDA Proposes New Framework for Regulating AI in Consumer Digital Health Tools

FDA Proposes New Framework for Regulating AI in Consumer Digital Health Tools

The U.S. Food and Drug Administration has issued draft guidance outlining marketing submission recommendations for a predetermined change control plan for artificial intelligence and machine learning-enabled device software functions as part of its oversight of consumer digital health tools. The document builds on the agency’s 2019 general wellness policy and the 2020 establishment of the Digital Health Center of Excellence. FDA guidance states that the agency applies a risk-based framework to AI-driven digital wellness and fitness applications rather than one based on the underlying technology.

Low-risk general wellness products that do not make claims about diagnosing, curing, mitigating, preventing, or treating disease fall outside FDA medical device authority. The 2023 guidance recommends that manufacturers outline anticipated modifications to AI/ML algorithms in premarket submissions when tools meet the definition of Software as a Medical Device.

What this means

The draft guidance clarifies how existing FDA policies on general wellness and AI/ML-enabled software apply to newer generative and adaptive AI systems in consumer fitness applications. It supplies illustrative examples of claims that trigger or avoid regulatory oversight and identifies documentation practices for algorithm updates. The materials indicate that many consumer-facing wellness tools are likely to remain outside FDA device regulation provided they confine themselves to low-risk general wellness uses, while those crossing into regulated territory are expected to include structured change control plans.

Criteria Distinguishing General Wellness Products from Medical Devices

The FDA’s 2019 General Wellness: Policy for Low Risk Devices defines when wellness and fitness products fall outside medical device authority [2]. Products intended only for general wellness use with low risk to users are not regulated as medical devices when they avoid disease-related claims [2].

FDA Approach to Adaptive AI/ML Algorithms

FDA guidance issued April 3, 2023 recommends that manufacturers submit predetermined change control plans for AI/ML-enabled device software functions [3]. The draft guidance addresses adaptive and generative AI systems by clarifying how existing SaMD policies apply to algorithm modifications in borderline wellness tools [3].

Developer Recommendations for Transparency and Risk Management

The FDA expects transparency, validation, and change management practices for AI/ML health technologies [3]. Manufacturers are directed to document how anticipated modifications to AI algorithms will be controlled, validated, and submitted when a tool qualifies as regulated software [3]. The Digital Health Center of Excellence, established September 22, 2020, serves as the hub for regulatory expertise on these digital health, software, AI, and wellness matters [1].

Relation to Prior FDA Digital Health and SaMD Policies

The current draft guidance extends the risk-based framework described in the 2019 general wellness policy and the 2020 Digital Health Center of Excellence announcement [1][2]. It applies the predetermined change control plan approach to consumer digital health tools that incorporate newer forms of adaptive AI while maintaining the distinction between general wellness products and regulated devices [3].

Limitations

Guidance remains in draft form and is subject to revision after public comment. It does not address data privacy, cybersecurity, or third-party AI supply-chain controls in depth. The document focuses on FDA jurisdiction and does not cover FTC or state-level consumer protection rules.