Supernova Technologies

I appreciate the opportunity to comment on this discussion paper. My comments address Questions 18, 19, 20, 24, and 25.

The paper’s proposed reliance on postmarket monitoring, and its proposed reliance on voluntary self reported information from foundation model developers, both depend on an assumption that has not been established: that the parties being monitored can reliably detect and report their own systems’ failures. This year’s pattern of AI safety incidents across multiple major AI developers, in which anomalous model behavior went undetected for weeks or months and was discovered by accident rather than through the developers’ own monitoring, suggests this assumption deserves scrutiny before it becomes load bearing for a regulatory framework.

On Foundation Model MAFs (Question 25): if developers with the strongest possible incentive to know their own models’ behavior have repeatedly failed to detect meaningful deviations, a voluntary model card filed by that same party carries the same structural limitation. I recommend CDRH require MAF filers to disclose their own internal detection methodology, not just claimed model behavior, so reviewers can assess the reliability of the process behind the claim.

On postmarket monitoring (Questions 18 and 19): a report of zero adverse events could mean the device performed safely, or that monitoring failed to observe the relevant behavior. These are not distinguishable from an absence of reports alone. I recommend sponsors be required to provide affirmative evidence of monitoring coverage, not just adverse event counts.

On degradation monitoring (Questions 19 and 24): a degrading device may not reliably signal its own decline through the same channel used to generate its outputs. I recommend degradation monitoring require at least one detection mechanism independent of the device’s own reporting pathway, such as periodic re benchmarking or sampled clinician review.

These recommendations support, rather than oppose, the competency based framework and reliance on postmarket monitoring. The goal is ensuring monitoring and disclosure mechanisms demonstrate their own reliability before being credited toward reduced premarket evidence.

Comment on Docket FDA-2026-N-7874
Considerations for the Regulation of Generative AI Enabled Medical Devices: Discussion Paper and Request for
Feedback

Submitted in response to Discussion Questions 18, 19, 20, 24, and 25

Introduction
I appreciate the opportunity to comment on this discussion paper. My comments focus on a single connected
concern that runs through Sections VI and VII: the paper's proposed reliance on postmarket monitoring, and its
proposed reliance on voluntary self reported information from foundation model developers, both depend on an
assumption that has not been established, namely that the parties being monitored can reliably detect and report their
own systems' failures. Evidence from this year's pattern of AI safety incidents across multiple major AI developers
suggests this assumption deserves close scrutiny before it becomes load bearing for a regulatory framework.

Point One: Voluntary Foundation Model Master Files (Question 25)
The paper directly asks what would make a voluntary Foundation Model MAF program sufficiently useful given
that model developers may have limited incentive to disclose safety relevant information. This is the right question
to ask, and the paper's own framing already identifies the core weakness.

Organizations with strong internal incentives to understand their own systems' behavior, including several leading
AI developers, have this year experienced real world incidents in which anomalous or unsafe model behavior went
undetected for extended periods, in some cases weeks or months, and was ultimately discovered through external
review or by accident rather than through the organizations' own monitoring processes. These were not
organizations lacking sophistication or motivation to monitor their systems. If developers with the strongest possible
incentive to know what their own models are doing have repeatedly failed to detect meaningful deviations in their
own systems, a voluntary model card or system card filed by that same party carries the same structural limitation: it
can only disclose what the filer already knows, and this year's pattern indicates that what a developer believes it
knows about its own system's behavior is not reliable.

I recommend CDRH consider requiring, as a condition of a Foundation Model MAF being referenced in a premarket
submission, that the filer disclose its own internal detection and monitoring methodology for the behaviors and
failure modes described in the card, so that reviewers can evaluate not just the claimed behavior of the model but the
reliability of the process that produced the claim.

Point Two: Postmarket Monitoring Completeness (Questions 18 and 19)
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.