← All 95 filings
Qiong Liu
Public / patientsMember of the publicFiled September 11, 20261,316 words · 1 attachmentFDA-2026-N-7874-0088
What they argued
RecovryAI’s one-line reading of the filing.
M2 and M4 from his section II (Q7, Q9): the benchmarking plus risk-proportionate clinical confirmation frame is right, conditioned on adding multi-turn history-completeness scoring against a pre-specified case-locked question list, same-transcript consistency metrics, instruction-lock guardrail tests, and shared item definitions with a hidden holdout and licensed review of decision gold (Q10, Q16). M5 from his summary ask 6 (Q22, Q24): re-run those measures when the foundation-model identifier or the prompt template changes. He states no position on device autonomy and no autonomy level. Type: he identifies as a medical-school graduate and software engineer, not licensed to practice, filing individually about a non-deployed personal prototype; no organisation, clinician licence, or academic affiliation is claimed, so public is the closest fit.
Themes it raises
5 of the 21 themes in the docket.
What makes a function high riskFDA Q1, Q2, Q5
Judging devices the way clinicians are credentialedFDA Q7, Q8
Whether benchmark results prove anythingFDA Q9, Q10, Q16
Controlling a device that keeps changingFDA Q22, Q23, Q24, Q25
Harm from an output that was not wrongFDA Q1, Q2
FDA questions it names
Questions this filing names by number.
Q1 · The two-axis risk frameworkQ5 · Multi-turn conversations that migrateQ6 · Care escalation functionsQ7 · The competency-based approachQ9 · The benchmarking structureQ10 · Benchmark contamination and saturationQ16 · Independent third partiesQ19 · Postmarket performance evaluationQ22 · Re-benchmarking after a modificationQ24 · Third-party foundation model changes
Across the five cross-cutting questions
RecovryAI’s reading of the whole filing. Silence is never counted as opposition.
Patient-facing autonomyShould FDA permit patient-facing AI to act with meaningful autonomy within a defined scope?
No position stated
Proportionate evidenceShould evidence requirements scale with clinical risk rather than a uniform high bar?
Supports with conditions
Postmarket relianceCan strong postmarket monitoring justify accepting more premarket uncertainty?
No position stated
Competency evaluationCan a device be evaluated on competency benchmarks and clinical confirmation against clinicians?
Supports with conditions
Change controlCan devices on third-party foundation models be maintained under pre-specified change control?
Supports with conditions
Autonomy acceptedThe highest level this filing accepts
Low-consequence work: Not stated
High-consequence work: Not stated
This filing was submitted by a member of the public. We publish our analysis here and link to the submission rather than reproducing it, and this page is not indexed by search engines.
Read the full submission on regulations.gov ↗