FDA Relaxes Clinical Decision Support and General Wellness Guidance: What It Means for Generative AI and Consumer Wearables
FDA Signals an Innovation-Friendly Approach to AI and Wellness Technologies
On January 6, 2026, the Food and Drug Administration (FDA) issued pivotal updates to its guidance on Clinical Decision Support Software (CDS) and General Wellness (GW) products. Taken together, these updates reflect a meaningful shift toward a less restrictive, more innovation-friendly regulatory environment for generative artificial intelligence (GAI) and consumer wearables.
These changes are not occurring in a vacuum. They align with broader Executive Branch efforts to reduce regulatory barriers to responsible AI adoption and to promote U.S. leadership in digital health innovation.
For digital health companies, the updated guidance provides clearer and more predictable pathways to market, particularly by expanding and clarifying the scope of FDA’s long-standing non-device “safe harbors.”
The Cures Act and FDA’s Evolving Approach to Non-Device Software
The 21st Century Cures Act of 2016 carved out important exceptions to what FDA considers a “medical device.” Among these are statutory exceptions for certain types of CDS software and general wellness products.
These exceptions have long served as an on-ramp for innovation, allowing digital health companies to test, deploy, and refine technologies in lower-risk, less-regulated contexts before pursuing regulated medical use cases. As digital health technologies have grown more sophisticated, FDA’s interpretation of these exceptions has evolved accordingly.
The January 2026 updates continue FDA’s effort to strike a balance between minimizing unnecessary regulatory friction and maintaining risk-based oversight, particularly as AI-driven tools become more prevalent in clinical and consumer settings.
Clinical Decision Support Software: When Is CDS a Medical Device?
Device CDS vs. Non-Device CDS: Why the Distinction Matters
For digital health companies, one threshold question often determines everything that follows: Is the software regulated as a medical device, or does it qualify as “non-device” Clinical Decision Support (CDS)?
That distinction matters because non-device CDS falls outside FDA’s medical device regulatory framework, allowing companies to bring products to market without premarket clearance, registration, or ongoing device compliance obligations. By contrast, device CDS is subject to FDA oversight, which may require a 510(k) clearance, De Novo authorization, or other regulatory controls depending on the software’s risk profile and functionality.
Under the 21st Century Cures Act, certain CDS software qualifies for a statutory exception from the medical device definition when it is designed to support, rather than replace, clinical decision-making. In general terms:
Non-Device CDS provides recommendations or insights for a healthcare practitioner, while preserving the practitioner’s ability to independently review the basis for those recommendations and exercise clinical judgment.
Device CDS crosses into regulated territory when the software effectively drives clinical decision-making, limits practitioner discretion, relies on opaque logic, or directly diagnoses or treats a patient without meaningful human review.
Historically, the line between device and non-device CDS has been narrow, highly fact-specific, and difficult to navigate — particularly for AI-driven tools. FDA’s updated CDS guidance meaningfully reshapes that line, especially for generative AI systems that produce a single, clinically appropriate recommendation for practitioner review.
The Four Core Criteria for Non-Device CDS Under the Cures Act
While the Cures Act initially created a statutory device exception for certain CDS software, subsequent FDA guidance has further defined the practical boundaries of that exception. To qualify as non-device CDS, software must meet four core criteria related to:
The type of data the software uses;
The nature of the software’s outputs;
The role of the healthcare practitioner in reviewing and applying those outputs; and
The level of transparency needed for the practitioner’s independent review.
Historically, these requirements significantly constrained the functionality permitted for non-device CDS, particularly for advanced AI-driven tools seeking to remain outside FDA regulation.
What Changed in FDA’s 2026 CDS Guidance?
The most impactful change in the updated guidance is FDA’s retreat from its long-standing position that non-device CDS must present multiple options or recommendations to a healthcare practitioner.
Previously, FDA viewed CDS software that presented a single preventive, diagnostic, or treatment option as raising heightened concerns about “automation bias,” often pushing such tools into regulated medical device territory.
Under the 2026 update, FDA adopts a more practical and risk-based approach. Software that presents a single clinically appropriate recommendation for practitioner review may still fall within the statutory definition of a medical device as a technical matter. However, FDA now intends to exercise enforcement discretion and not apply medical device regulatory requirements when:
There is only one clinically appropriate option to recommend; and
The software otherwise satisfies all other requirements for the non-device CDS exception, including meaningful practitioner review and transparency.
In other words, single-output CDS is no longer automatically disqualifying for non-device treatment — provided the practitioner remains firmly in the decision-making loop.
FDA Examples That Benefit Generative AI and Clinical Copilots
FDA’s revised guidance includes illustrative examples of CDS functionality that may qualify for enforcement discretion, including software that:
Predicts a patient’s risk of future health events for practitioner consideration;
Generates recommended treatment plans based on patient data for practitioner review, revision, and approval;
Produces a differential diagnosis based on symptoms, vitals, and lab values for practitioner review and finalization; and
Analyzes clinical findings and generates draft clinical summaries, including diagnostic recommendations, for practitioner review and revision.
Although application of these examples remains highly fact-specific, the guidance creates a practical go-to-market pathway for GAI-driven decision support tools, including clinical copilots, assistants, and chat-based workflows that generate a single best recommendation while preserving clinician judgment.
General Wellness Products: FDA Reopens the Door for Consumer Wearables
When Are Wearables Still Outside FDA Regulation?
General wellness products have long fallen outside FDA regulation when limited to promoting healthy lifestyles. However, advances in consumer wearables have increasingly blurred the line between wellness tracking and medical use.
That ambiguity intensified following recent FDA enforcement actions against wearable companies that historically positioned their products as non-medical — most notably FDA’s warning letter to WHOOP regarding its Blood Pressure Insights feature. In that letter, FDA asserted that blood pressure measurement is inherently tied to diagnosing hypertension or hypotension, raising concerns that many consumer wearables could be pulled into medical device regulation by default.
What the Updated General Wellness Guidance Clarifies
FDA’s updated General Wellness Guidance meaningfully walks back this position and provides long-awaited clarity. FDA now confirms that non-invasive sensing used to estimate physiological parameters — such as blood pressure, oxygen saturation, or glucose — does not automatically make a product a medical device.
Instead, such products may qualify as general wellness products when:
Outputs are intended solely for non-medical wellness purposes;
The product is non-invasive and not implanted;
The product presents an appropriately low risk to users;
The product is not intended to substitute for an FDA-cleared or approved medical device;
The product does not prompt or guide clinical management of a medical condition; and
The output does not mimic clinical values unless appropriately validated.
FDA also clarifies that wellness products may notify users to seek professional medical evaluation when data falls outside normal ranges, so long as those notifications avoid diagnosing, naming diseases, or providing medical interpretation.
This clarification mirrors the CDS guidance’s broader theme of preserving human judgment and avoiding automation or medical direction.
Key Takeaways: How FDA’s Updated Guidance Affects Digital Health Companies
A Clearer Path for Single-Output Generative AI CDS
FDA’s updated CDS guidance expands the universe of tools that can remain non-device CDS, even when presenting a single clinically appropriate recommendation, provided practitioner review and transparency requirements are met. This unlocks use cases that previously forced AI developers into regulated device pathways.
Reduced Regulatory Friction for AI Clinical Copilots
AI tools that summarize clinical findings, propose diagnoses, or generate draft documentation now have a clearer path to market without device clearance — if properly designed and positioned.
Expanded Safe Harbor for Consumer Wearables
Wearables that estimate physiological metrics for non-medical wellness use can remain outside FDA regulation when outputs, claims, and user messaging are carefully constrained.
Notifications from Wellness Wearables Are No Longer Automatic Red Flags
FDA now permits wellness products to prompt users to seek medical evaluation based on abnormal data, so long as strict guardrails are followed.
Policy Alignment with Broader AI Initiatives
These updates reflect a broader federal policy trend toward reducing regulatory barriers to responsible AI adoption in healthcare.
Design and Positioning Still Matter
FDA’s examples are illustrative, not determinative. Small changes in product design, labeling, or workflow can materially change regulatory classification. While this updated guidance expands opportunity, it does not eliminate the need for careful, feature-level regulatory analysis.
A Wider Path to Market, but Guardrails Still Required
For digital health developers, FDA’s updated guidance marks a meaningful recalibration of risk tolerance, particularly for AI-driven software and consumer-facing health/wellness technologies. Companies that understand and design with FDA’s guidance in mind may now have a faster, compliant path to market.
As always when it comes to healthcare, the devil is in the details. At Nixon Law Group, we help digital health and SaMD companies navigate FDA’s evolving regulatory landscape, from early classification strategy to 510(k) submissions, post market compliance, and marketing review. If you are seeking clarity on how the updated guidance affects your product or business model, please contact us. We’re here to help!