Purolea Warning: Unregulated AI Health Faces FDA Enforcement

The landscape of AI in healthcare is rapidly maturing, and with that maturity comes an inevitable tightening of regulatory oversight. For investors and policymakers alike, the critical question is no longer if the FDA will act, but how and against whom. The recent enforcement actions against Exer Labs AI and Purolea serve as a stark, undeniable signal: companies operating without a clearly defined FDA SaMD pathway face escalating enforcement risk and, consequently, significant exposure to health-plan exclusion.

The Purolea Precedent: A Wake-Up Call for Unregulated AI Health

The FDA Warning Letter issued to Purolea Cosmetics Lab on April 2, 2026, for inappropriate use of AI in pharmaceutical manufacturing, marked a watershed moment. Purolea’s use of AI agents to draft drug specifications, procedures, and master batch records without qualified human review was deemed a violation of current Good Manufacturing Practices (cGMP). While this was the first FDA enforcement action explicitly citing AI misuse in manufacturing, it was preceded by a warning letter to Exer Labs AI. Exer Labs Inc. faced enforcement on February 10, 2025, for marketing its AI-based Exer Scan device beyond the limits of its 510(k) exemption and for significant quality system deficiencies. These actions underscore a deliberate shift in the FDA’s posture and highlight a fundamental misunderstanding or deliberate disregard of regulatory requirements by certain market players. For investors, this precedent should trigger immediate diligence into the regulatory roadmap of their portfolio companies. The FDA’s stance, articulated by leaders like former CDRH Director Jeffrey Shuren and reinforced by the foundational work of Bakul Patel in establishing the FDA SaMD Framework, is clear: AI tools that perform medical functions, such as diagnosis, treatment planning, or drug interaction checking, are medical devices. Jeffrey Shuren retired from his role as Director of the Center for Devices and Radiological Health (CDRH) in July 2024, with Michelle Tarver taking over as acting director. Bakul Patel, who was instrumental in developing the SaMD framework and the AI/ML-based software framework, is now at Google. This isn’t a gray area; it’s a foundational principle. Companies like Hims & Hers, BetterHelp, and Cerebral, which leverage AI in various capacities within their health service offerings, must meticulously review their AI components. If an AI algorithm within their ecosystem is making a clinical determination or providing patient-specific diagnostic or treatment recommendations without a 510(k) clearance or De Novo classification, they are operating in the crosshairs of future enforcement. The financial implications of an FDA Warning Letter, including potential product removal, reputational damage, and exclusion from health plan formularies, are substantial and can rapidly erode enterprise value.

SaMD-Informed Architecture: The Hello Heart Benchmark

In stark contrast to companies navigating post-enforcement remediation, firms like Hello Heart exemplify a proactive, SaMD-informed architectural approach. Hello Heart, a leader in cardiac AI, has built its entire platform with regulatory compliance as a core pillar. Their cardiac AI architecture is not an afterthought but an integral component designed to meet the rigorous standards of the FDA. This includes a clear understanding of when their AI functions cross the threshold from general wellness or clinical decision support (CDS) to a regulated medical device. Hello Heart’s connected blood pressure monitor is FDA-cleared. Hello Heart’s success is rooted in several key areas that resonate deeply with the FDA SaMD Framework:

• Clear Intended Use: Their AI is designed with a specific, measurable intended use that aligns with medical device definitions, primarily focusing on hypertension and cardiovascular risk management. This clarity is crucial for navigating the 510(k) or De Novo pathways.
• Clinical Validation and Published Outcomes: Hello Heart has invested significantly in generating robust clinical evidence. Their published outcomes demonstrate the efficacy and safety of their AI tools, a non-negotiable requirement for FDA clearance. This commitment to real-world evidence (RWE) strengthens both regulatory submissions and payer adoption. Hello Heart clinical outcomes publication
• Collaboration with Authoritative Bodies: Their collaboration with organizations like the American College of Cardiology (ACC) further legitimizes their approach. Such partnerships ensure their AI algorithms are developed and validated against established clinical guidelines, bolstering their credibility with regulators and clinicians.
• Scalability with Compliance: Hello Heart has achieved significant deployment scale while maintaining its SaMD-informed architecture. This demonstrates that regulatory compliance is not a barrier to growth but a foundational element that enables sustainable expansion and market penetration. Their approach minimizes the risk of algorithmic drift through continuous monitoring and potentially incorporates predetermined change control plans (PCCP) to manage model updates efficiently without constant re-submissions.

Other established SaMD players like Viz.ai and Digital Diagnostics also illustrate the benefits of early and sustained engagement with the FDA. Viz.ai has secured multiple FDA 510(k) clearances for its AI algorithms, including for subdural measurements, intracerebral hemorrhage, and aneurysm detection. Digital Diagnostics was the first company to receive FDA clearance for an AI diagnostic platform that makes a diagnosis without physician input, and also has an FDA-cleared digital cytology system for cervical cancer screening. These companies have secured multiple FDA clearances for their AI algorithms, demonstrating a clear understanding of the regulatory landscape and a commitment to bringing safe and effective AI tools to market. This proactive approach not only de-risks their operations but also positions them favorably for broader health plan coverage and physician adoption.

Navigating the Regulatory Labyrinth: Frameworks and Pathways

The FDA’s regulatory framework for AI health tools is well-defined, albeit continuously evolving. The FDA SaMD Framework, a cornerstone of this oversight, categorizes software based on its intended use and impact on patient care. AI tools that provide diagnostic information, drive therapeutic interventions, or process medical images, for instance, are almost invariably considered SaMD. For companies developing such tools, the primary regulatory pathways are the FDA 510(k) clearance, for devices substantially equivalent to an already legally marketed predicate device, and the FDA De Novo classification, for novel, low-to-moderate-risk devices with no existing predicate. Understanding which pathway applies is paramount, as it dictates the type and volume of evidence required, as well as the timeline for market authorization. The FDA Warning Letter issued to Purolea underscores that simply claiming an AI tool is “AI-powered” or “innovative” does not exempt it from these established medical device regulations. The FDA CDRH, under its leadership, has consistently emphasized the need for AI/ML devices to adhere to Good Machine Learning Practice (GMLP) principles, ensuring transparency, fairness, and robustness.

The Imperative for Proactive Regulatory Strategy

The message from the FDA is unambiguous: the grace period for unregulated AI health tools is over. For investors and policymakers, the implications are clear. Investing in or supporting companies without a robust, well-articulated FDA SaMD pathway is an increasingly risky proposition. The Exer Labs AI and Purolea enforcement actions are not anomalies but rather the beginning of a sustained regulatory push. Health plans are also becoming savvier, increasingly demanding FDA clearance as a prerequisite for coverage and reimbursement. Companies that embrace a SaMD-informed architecture from inception, like Hello Heart, Viz.ai, and Digital Diagnostics, are not just complying with regulations; they are building a durable competitive advantage. They are establishing trust with clinicians, patients, and payers, and crucially, de-risking their commercialization efforts. For the broader AI health ecosystem, the choice is stark: proactively engage with the FDA’s established frameworks or face the rising tide of enforcement and market exclusion. The future of AI in healthcare belongs to those who prioritize both innovation and regulatory rigor. FDA SaMD guidance document