Can greater premarket uncertainty about a GenAI device’s benefit-risk profile be accepted through greater reliance on postmarket monitoring?

The question as posed CDRH is considering whether it may be appropriate to accept greater premarket uncertainty regarding a GenAI-enabled device's benefit-risk profile through greater reliance on postmarket monitoring. Under what conditions might such an approach be appropriate, and what characteristics of a monitoring program would need to be in place to justify reduced premarket evidence? Are there device types or risk profiles for which this approach would not be appropriate? Summary of position We support the trade in principle. A total product lifecycle approach is the right instinct for devices whose behavior is open-ended and whose models change after clearance. Our comment is about the condition on which the trade depends, and about one class of device for which we recommend it not be available at all. • A reduction in premarket evidence is only a trade if the monitoring program can detect the failure the premarket evidence would have caught. Where it cannot, nothing has been exchanged — certainty has been given up for coverage that does not exist. We recommend the reduction be granted against demonstrated detection capability, never against the existence of a monitoring program. • We have measured a failure mode that a periodic monitoring program cannot see by construction. It occurs within a single session and resets at the session boundary, so every scheduled reassessment measures the healthy state. A program built on periodic re-benchmarking should not be able to fund a premarket reduction for any failure mode of that shape. • Monitoring that reads the device's own account of its work inherits the device's errors. If premarket evidence is reduced on the strength of postmarket monitoring, and that monitoring is a self-report, the reduction rests on the device vouching for itself. • We recommend the approach be unavailable where the intended user cannot self-verify or self-report, where the device operates without connectivity, and where the harm completes inside the detection interval. Each Docket FDA-2026-N-7874 · The Christman AI Project and Robotics Division Page 1 of these describes the assistive and communication devices we build, and each removes a safeguard the trade quietly assumes is present. 1. The unstated premise in the trade Accepting greater premarket uncertainty in exchange for postmarket monitoring is a sound instrument, and it is already how much of the device program works. But the exchange carries a premise that is rarely written down: that the residual uncertainty is of a kind postmarket monitoring can resolve. For most device attributes that premise holds. A durability question, a failure rate, a rare adverse event — these accumulate in the field and become visible with time and volume. Postmarket surveillance is the correct instrument because the signal is statistical and time reveals it. The failure modes we have measured in generative systems are not of that kind, and time does not reveal them. They are individually invisible, they do not aggregate into a rate, and they are erased by the very act of reassessment. We set out the measurement below, because the recommendation follows from it rather than from a position. 2. The measurement, and why it defeats a periodic program On 2026-09-03 we made three recordings across seventy-four minutes of continuous work on one unchanged audio interface, with no configuration change between them. Input carrying no live signal accounted for 7.1 percent of the first file, 19.1 percent of the second, and 42.2 percent of the third. The proportion of lost input roughly doubled between each recording. The consequence for Question 18 is structural rather than a matter of rigor. A re-benchmark is a fresh session. A reassessment run at any point on any day would have opened a new session and measured something close to the 7.1 percent state. The degradation is a function of elapsed time within a session, and a monitoring cadence measured in weeks or quarters cannot observe a quantity that resets in minutes. A separate recorded session on 2026-09-04, eleven minutes twenty-four seconds, showed the second half of the problem. The device produced false statements in three separate turns while its presentation improved steadily across the session — by the ninth minute it was citing governing rules by name, disclosing source ages, and correcting itself unprompted. A sample drawn late in that trajectory scores the device as more disciplined than a sample drawn early, while the error rate is unchanged. A monitoring program that samples will report the better number, and will report it in good faith. We state the boundary on this evidence in Section 6. It establishes that these failure modes occur and can be measured. It establishes nothing about how often, and we offer no rate. 3. Conditions we recommend attaching to any reduction Responsive to the first and second parts of Question 18. We propose three conditions, each stated so that a reviewer can de

R2. Do not trade premarket evidence for postmarket monitoring outside the lowest-risk quadrant (Questions 18, 19, and 21) This follows from S1 and S2. Question 18 asks whether CDRH should "accept greater premarket uncertainty regarding a GenAI-enabled device's benefit-risk profile through greater reliance on postmarket monitoring." My answer is a qualified no. The framework in S1 exists precisely to identify which functions can tolerate uncertainty. If a function sits in the lower-left quadrant — non-directive information with limited consequences — then reduced premarket evidence with defined monitoring is a reasonable proportionality judgment. For anything action-directing or action-taking, or anything with moderate or severe consequences, accepting premarket uncertainty transfers risk from the sponsor, who chose to deploy, to the patient, who did not. Docket No. FDA-2026-N-7874 — Individual comment — Page 3 Postmarket surveillance is also structurally weaker than the paper implies. Existing device adverse event reporting depends on passive reporting and is known to under-capture harm. GenAI failure modes — confabulation, cumulative scope drift, over-reassurance, silent degradation after an upstream model change — have no established reporting taxonomy, no established detection method, and no natural reporter, since the patient may never learn that an output was wrong. Before postmarket monitoring can justify reduced premarket evidence, CDRH would need at least the following in place: a reporting taxonomy specific to GenAI failure modes; prespecified, publicly disclosed performance thresholds with a defined cadence of re-benchmarking; a patient-facing channel for reporting suspected incorrect outputs; and public reporting of monitoring results, not just submission to FDA. On Question 21, I would caution against "shared ecosystem responsibility" becoming a mechanism for diffusing accountability. Clinicians, institutions, and professional societies can contribute signal. They should not absorb liability. The sponsor is the manufacturer and remains responsible for the device's performance across the total product life cycle, including for the behavior of components it licenses from others.

Question 18: Conditions for relying on postmarket evidence. Greater reliance on postmarket evidence may be appropriate when the manufacturer supports the initial benefit-risk determination, identifies the remaining uncertainty, and demonstrates a feasible plan to resolve it. The justification should consider the available alternatives and the consequences of delaying access. The manufacturer should define the outcomes to collect, the evidence target and deadline, and limits on exposure while uncertainty remains. Reviewers should consider severity and reversibility of harm, cumulative patient exposure, and the ability to detect rare or delayed failures. The manufacturer should specify criteria for restricting use if evidence collection falls behind or results exceed the prespecified risk limits. Monitoring alone is inadequate for a failure that could cause serious harm before detection and effective intervention. The manufacturer should compare time to harm with the combined time needed to detect, investigate, and contain the failure. It should account for unavailable reviewers and supplier delays. Such a failure requires preventive controls, restricted functionality, or another justified safeguard before deployment.

Question 18: More postmarket reliance CDRH can accept more premarket uncertainty only when the manufacturer can detect and control the remaining risk. The device should have: A limited initial deployment. A safe fallback process. Continuous version records. Defined monitoring measures. Defined alert limits. Named persons who review alerts. A rapid stop or rollback method. A plan to collect the missing evidence. This approach is not suitable when harm is irreversible or time-critical. It is also not suitable when the manufacturer cannot detect failure before harm.

3.7 Question 18 - Premarket uncertainty and postmarket evidence The FDA asks whether some additional premarket uncertainty might be acceptable where strong postmarket monitoring exists (FDA, 2026a). Sitora considers this potentially appropriate for selected lower-risk functions where outputs are reversible, human review is effective, the system can be rapidly rolled back, audit logging is comprehensive and autonomy is limited. A progressive deployment model could move through controlled technical validation, shadow-mode clinical evaluation, supervised deployment, broader monitored deployment and, only where separately justified, defined autonomous permissions. Greater flexibility should not be presumed for functions that are irreversible, immediately life-critical or otherwise high consequence.

Question 18 (reliance on postmarket monitoring). Accepting greater premarket uncertainty in exchange for postmarket monitoring is not appropriate for functions that produce permanent implants. Monitoring can detect a problem in this setting; it cannot remediate one without surgery. Where the harm is irreversible, premarket assurance of the manufacturing route should not be traded against surveillance.

Discussion Question 18 – Premarket vs. Postmarket Activities In principle, increased reliance on postmarket monitoring can be appropriate, but it should not create a general trade-off whereby stronger postmarket monitoring automatically permits less premarket evidence. The Discussion Paper specifically frames this question in terms of uncertainty regarding the device's benefit-risk profile, which ultimately remains a product-specific determination. The acceptable degree of premarket uncertainty should primarily depend on the device's criticality (as introduced in the feedback to Sections IV and V) and on whether emerging risks can be detected and controlled before clinically significant harm occurs. Relevant considerations include: • severity and reversibility of potential harm; • detectability and latency of emerging failures; • degree of autonomy and effectiveness of human or technical safeguards; • number and rate of patients exposed; • ability to restrict use, suspend operation, or rapidly roll back changes; and • stability of the underlying model and deployment environment. Higher-criticality devices should generally require greater premarket assurance and more intensive postmarket monitoring. Increased postmarket monitoring should be able to compensate for premarket uncertainty only where the residual uncertainty is sufficiently observable, reversible, and controllable. Postmarket monitoring should additionally verify continued compliance with the assumptions used in the Criticality Stratification, including the actual effectiveness of human oversight, scope limitations, escalation procedures, and other safeguards.

Question 18: Accepting greater premarket uncertainty. A pre-/post-market rebalancing is ill-suited for autonomous functions that take actions with severe consequences; furthermore, for functions where users cannot detect errors, post-market signals would only become apparent after harm has occurred, a situation that pre-market evidence is specifically intended to prevent. Incorporating these conditions into the FDA’s existing guidance on benefit-risk uncertainty would maintain a doctrinally consistent approach rather than an exceptional one.

Question 18 - Greater premarket uncertainty supported by postmarket monitoring Greater reliance on postmarket monitoring may be appropriate when the residual uncertainty is measurable, monitoring can detect clinically meaningful degradation before unacceptable harm accumulates, intervention or rollback is feasible, and the postmarket evidence system is sufficiently complete to support attribution and corrective action. [1] Reduced premarket evidence is less appropriate where the function is fully autonomous, consequences are severe or irreversible, failures may be difficult to detect promptly, or the system lacks reliable postmarket observability. A monitoring plan should not be used to compensate for a premarket evidence gap that could expose patients to uncharacterized high-consequence risk before the monitoring system can react.

Question 18 — Trading premarket certainty for postmarket monitoring Yes — but only inside a defined risk envelope, with a real safety net underneath. For low- and moderate-risk systems with reversible harms and continuously measurable performance, FDA can reasonably accept more premarket uncertainty in exchange for real-time monitoring and clear intervention thresholds. For high- consequence autonomous functions, postmarket monitoring should supplement adequate premarket evidence, not substitute for it. The more uncertainty accepted going in, the stronger the detection-and-correction mechanism required coming out.

Question 18. Greater postmarket reliance is appropriate only when monitoring is timely, actionable, and enforceably connected to risk controls. Reduced premarket certainty should not be justified by passive dashboards. The sponsor should prespecify signals, thresholds, review cadence, escalation paths, authority to pause or restrict functionality, corrective-action timelines, and evidence that monitoring coverage is adequate. High-consequence autonomous actions, rapid time-to-harm, weak reversibility, or inadequate interruption mechanisms may make reduced premarket evidence inappropriate. Postmarket observability should be proportionate and privacy preserving. Routine evidence can emphasize configuration identifiers, policy results, uncertainty ranges, safety-control Public Comment — Generative AI-Enabled Medical Devices VIVASECURIS | FDA-2026-N-7874 interventions, device state, clinician response, and aggregate performance rather than indiscriminate retention of complete prompts or identifiable clinical records. A graduated model can support minimal operational metadata during normal use, controlled correlation when a risk signal emerges, and narrowly authorized access to case-level evidence for investigation. Additional consideration related to Questions 18 and 21. Data collected or created by GenAI requires an explicit lifecycle, not a generic privacy statement. GenAI systems can create new sensitive information even when the source record is already governed. Conversations, inferred symptoms or diagnoses, risk classifications, summaries, embeddings, retrieved context, agent memory, tool results, feedback, safety interventions, and audit records may reveal or amplify clinically significant facts. For data practices materially affecting device safety or effectiveness and within FDA's authority, FDA should ask sponsors to identify each data class, why it is necessary, where it is stored, who can access it, which downstream systems receive it, how long it persists, and how it is corrected, retained, archived, or deleted.  Purpose limitation and minimization: collect and derive only the information necessary for the authorized function, safety monitoring, and legally required records.  Separation: distinguish patient content, operational state, model context, long-term memory, quality evidence, cybersecurity evidence, and de-identified aggregate metrics so each can receive appropriate access and retention controls.  Identity and authorization: bind every read, write, inference, retrieval, export, and deletion to an authenticated human or machine identity, permitted purpose, patient scope, and time-bounded authority.  Retention and deletion: define limits for prompts, outputs, caches, embeddings, memory, backups, replicas, and supplier-held copies; verify deletion or justified archival rather than relying on user-interface disappearance, subject to applicable record-retention, investigation, litigation-hold, and data-integrity obligations.  Correction and provenance: preserve the source and transformation history of derived information and propagate corrections so an erroneous inference does not become durable clinical truth across systems.  Secondary use: do not silently use patient interactions, clinical content, or safety investigations for model training, product improvement, or unrelated analytics. Any authorized reuse should be separately governed, transparent, and technically enforceable.  Supplier boundaries: contracts and technical controls should address model-provider retention, human review, abuse monitoring, support access, regional storage, subprocessors, incident response, and termination or migration. Privacy and safety should not be framed as opposites. A system can preserve enough attributable evidence to investigate a serious failure without retaining every raw conversation indefinitely. Sponsors should justify the minimum evidence needed for causal reconstruction, protect content separately from metadata, and support controlled re-identification only when authorized for patient care, safety investigation, or regulatory obligations. Public Comment — Generative AI-Enabled Medical Devices VIVASECURIS | FDA-2026-N-7874

Question 18 Accepting greater premarket uncertainty in exchange for postmarket monitoring Drug safety has operated this trade-off for decades. Premarket exposure is limited by trial size and duration, and much of what is eventually known about a product's safety profile is learned after approval through spontaneous reporting and postmarketing surveillance. The experience is directly relevant here, and it is not uniformly encouraging. Three lessons seem to me transferable. First, postmarket systems detect what they were designed to detect. They are structurally weakest at identifying harms nobody specified in advance which is precisely the category that open- ended generative outputs are most likely to produce. A monitoring programme built around prespecified benchmarks will measure prespecified failure modes well and novel ones poorly. Second, the sensitivity of a postmarket system is governed by the quality of its individual records, not by their number. Aggregate analysis inherits every weakness of the case-level data beneath it. Where records are incomplete, internally inconsistent, or coded variably, signal detection degrades in ways that are difficult to see from the aggregate. Third, and most importantly for CDRH's question, quality systems tend to measure what is easy to count. Timeliness is easy to count; accuracy and completeness are not. A postmarket programme that reports on cadence adherence while never characterising its own detection capability can look healthy indefinitely while detecting very little. If premarket uncertainty is to be accepted in exchange for postmarket monitoring, I would suggest the condition should not be that a monitoring plan exists, but that its detection capability has been characterised. Specifically, before the trade is accepted, a sponsor might be expected to state: • which classes of failure the monitoring programme is designed to detect; • which classes it is not designed to detect, stated affirmatively rather than left as silence; • the expected interval between an occurrence and its detection; and • how each of those claims will be verified in operation rather than asserted at authorisation. A programme that cannot say what it would fail to notice is not yet a substitute for premarket evidence.

Question 18 — Can stronger postmarket monitoring justify greater premarket uncertainty? Response Sometimes, but only within reasonable limits. Generative and agentic AI are inherently dynamic, and some performance characteristics will not be fully visible until systems interact with real patients, clinicians, environments, and workflows. Deborah “Nurse Deb” Ault | Response to FDA Discussion Paper | Page 12 FDA-2026-N-7874 | Generative AI-Enabled Medical Devices Postmarket monitoring is therefore essential. But: Patients must not unknowingly become the clinical trial. The amount of acceptable premarket uncertainty should shrink as:  potential harm becomes more serious;  the harm becomes less reversible;  decisions become more time-sensitive;  the AI becomes more autonomous;  qualified human review becomes less immediate;  the system interacts directly with patients;  the patient population becomes more vulnerable; or  the AI can initiate consequential actions rather than merely provide information. A system capable of influencing whether someone seeks emergency care should not enter broad clinical use with the same tolerance for uncertainty as a low-risk administrative tool. Similarly, a system capable of autonomously denying, delaying, redirecting, or changing care requires substantially stronger evidence than one whose output is merely advisory and independently reviewed by a qualified clinician. Innovation speed is valuable. It is not a substitute for adequate evidence.

Question 18 — Trading premarket evidence for postmarket monitoring This is an appealing proposition and it may be the right direction, but I want to register a concern grounded in how postmarket commitments have historically performed. The general pattern across regulated products is that postmarket obligations are less reliably fulfilled than premarket ones — completion is slower, enforcement is resource-intensive, and passive surveillance systems depend on reporting behavior that is uneven. I do not have current verified figures on postmarket study completion rates or software-related adverse event reporting rates, and I would encourage CDRH to publish its own data on this before relying on postmarket monitoring as a substitute for premarket evidence. If the 11 of 19 Docket No. FDA-2026-N-7874 data show that postmarket obligations for software devices are reliably met, that materially strengthens the case; if they do not, the paper’s proposal transfers risk to patients in exchange for an obligation that may not be discharged. Conditions I would consider necessary: • The monitoring program must be an enforceable condition of authorization, with prespecified triggers and prespecified consequences — including suspension of marketing — that attach automatically rather than through a discretionary enforcement decision. A commitment is not a control. • The sponsor must demonstrate technical capability to observe the endpoint. Many GenAI devices have no linkage between an output and any recorded clinical outcome. Where the sponsor cannot show a data pathway from output to observable consequence, the monitoring program is aspirational and cannot justify reduced premarket evidence. This should be an explicit gating question. • Detection latency must be shorter than harm accumulation. Where a degradation would produce harm faster than the monitoring cadence could detect it, the trade is unavailable regardless of program quality. • Irreversibility forecloses the trade. Where the harm from an incorrect output cannot be undone, postmarket detection is not a mitigation — it is an audit of damage already done. I recommend CDRH state affirmatively that reduced premarket evidence is not available for devices in the high-consequence, high-activity region of the risk grid, so that the flexibility is bounded on its face.

Question 18 asks what characteristics of a monitoring program would need to be in place to justify reduced premarket evidence. I recommend that any postmarket monitoring program accepted as a basis for reduced premarket evidence be required to demonstrate the completeness of its own coverage, not merely the presence of a monitoring process. A monitoring program that reports zero adverse events could mean the device performed safely, or it could mean the monitoring failed to observe the relevant behavior. These two states are not distinguishable from the absence of a report alone, yet a regulatory framework built around greater reliance on postmarket monitoring needs to distinguish between them. I recommend that sponsors be required to provide affirmative evidence of monitoring coverage, for example documentation of what proportion of real world interactions were captured by the monitoring system and what categories of device behavior fall outside its observable scope, alongside any adverse event reporting. Point Three: Independent Detection of Performance Degradation (Questions 19 and 24) The paper names performance degradation over time as one of the core risks of GenAI enabled devices, and proposes performance degradation monitoring as a postmarket approach. I recommend that CDRH require degradation detection mechanisms to operate independently of the deployed device's own output or reporting pathway. A device that is degrading, whether due to a change in the underlying foundation model, drift in the input population, or another cause, may not reliably signal its own decline through the same channel used to generate its outputs, particularly for GenAI enabled devices where a single failure mode can affect both the device's clinical output and its own self assessment of that output. This concern is closely related to Question 24, regarding changes initiated by a third party foundation model developer rather than the device manufacturer. In both cases, the manufacturer's ability to detect a problem depends on a signal that is generated or mediated by the same system that may be experiencing the problem. I recommend that degradation monitoring specifications require at least one detection mechanism, such as periodic re benchmarking against an independent reference set or sampled clinician review as described in Section VI.A, that does not depend on the device's own reporting of its performance. Conclusion The concerns raised above are not arguments against the competency based framework or against reliance on postmarket monitoring generally. Postmarket monitoring, done well, is likely necessary given the practical limits of premarket testing for open ended systems described elsewhere in this paper. My recommendation is narrower: that CDRH require monitoring and disclosure mechanisms to demonstrate their own reliability and completeness as a condition of being credited toward reduced premarket evidence, rather than treating the existence of a monitoring program or a voluntary disclosure mechanism as sufficient on its own. Given this year's demonstrated pattern of AI developers failing to detect problems in their own systems despite strong incentives to do so, this distinction is likely to matter in practice. Thank you for the opportunity to comment.

Question 18: postmarket-weighted evidence. I support accepting greater premarket uncertainty in exchange for postmarket monitoring, under conditions: a prespecified monitoring plan with quantitative thresholds and defined triggers; automated performance and drift surveillance with sampled adjudication by independent clinicians; a transparent reporting cadence; and demonstrated rollback capability. A deployed conversational device generates more decision-relevant performance evidence in a month of instrumented use than any premarket sample can contain. A regulatory posture that credits well-built monitoring will also push manufacturers to build the instrumentation, which improves safety independent of any submission.

FDA Question 18 - Greater postmarket reliance / premarket uncertainty Trace ID. TR-Q18 | FDA Q18; Sec. VI.D; App. B; pp. 21-22 / 29-30 BCR response. Allow greater postmarket reliance only when residuals are bounded, detectability is proven, monitoring is sensitive, intervention precedes unacceptable harm, and actions are reversible or otherwise controlled. Do not defer catastrophic, irreversible, or poorly observable risk. BCR rule basis. BCR-R10,R12,R14,R15,R16 Solution-stack link. S4,S12,S13 Closure evidence. Detection sensitivity, monitoring latency, response time, reversibility and harm-time analysis Pass / re-open. Detection + response demonstrably precede unacceptable harm for deferred residuals Re-open when: Monitoring capability, harm latency, or action reversibility changes.

Response to Question 18: Accepting Greater Premarket Uncertainty — A Sound Trade The proposal to accept greater premarket uncertainty in exchange for stronger, more rigorous postmarket monitoring is conceptually sound and practically necessary for generative AI devices. The nature of large language model and generative AI performance — inherently probabilistic, contextually sensitive, and potentially subject to performance drift as clinical language and practice norms evolve — makes it impossible to fully characterize safety and efficacy through premarket evaluation alone. Real-world performance data, collected systematically under defined conditions, will always provide the most meaningful signal. WHM endorses this general principle with one critical condition: the postmarket monitoring plan must be fully prespecified at the time of market authorization. It cannot be left to post- clearance negotiation. The monitoring plan — including performance metrics, data collection methodology, reporting cadence, triggering events for reassessment, and defined performance thresholds that would trigger mandatory regulatory action — must be a legally enforceable commitment, not an aspirational document. Manufacturers who commit to robust postmarket monitoring should receive meaningful premarket flexibility; those who cannot commit to rigorous monitoring should face correspondingly rigorous premarket standards.