Prevention for LongevityInstitutionIntelligenceStandardWorkAboutDiscuss an opportunity

Evidence Explorer · EV-009

AI real-world performance

In a 2025 discussion-only request for public comment, FDA asked about post-deployment AI-device performance and drift, evaluation infrastructure, clinical outcomes and user feedback, and human-AI interaction.

Primary/authoritativeHistoricalFederal policy/science discussion2025-09-30

What this evidence supports

In a 2025 discussion-only request for public comment, FDA asked about post-deployment AI-device performance and drift, evaluation infrastructure, clinical outcomes and user feedback, and human-AI interaction.

What it does not establish

The request's comment period ended December 1, 2025. It is discussion-only—not draft or final guidance, an adopted framework, an enforceable requirement, or a policy change.

Answer-ready summary

FDA's September 2025 discussion-only request asked about lifecycle monitoring for AI-enabled medical-device performance; comments closed December 1, 2025, and the request is not guidance or policy.

Record details

Decision priority
Historical
Current status
Closed request for public comment · discussion only · not draft/final guidance or policy
Evidence tracks
Technologies/care models · Human outcomes
Source authority
Primary/authoritative
Source organization
FDA
Evidence design
Federal policy/science discussion
Applicability
Applicable to decisions about the stated program, policy, standard, or development context; not evidence of clinical benefit beyond the cited scope.
Confidence for the stated claim
High
Freshness
Historical
Record date
2025-09-30
Source updated
2025-09-30
Last verified
2026-09-03
Review due
On material source change; preserve as a dated record.
Version
1.1
Revision state
Version 1.1 corrects the record date from August 8, 2026 to the source publication date of September 30, 2025, recasts the item as a closed discussion-only request, and records the September 3, 2026 re-verification.

Evidence record fields

Claim supported
In a 2025 discussion-only request for public comment, FDA asked about post-deployment AI-device performance and drift, evaluation infrastructure, clinical outcomes and user feedback, and human-AI interaction.
Source type and design
Federal policy/science discussion
Population and setting
Not a participant-level clinical-study population in this source type. The program, policy, regulatory, standards, or research context described by the cited source.
Intervention/exposure and comparator
Not evaluated as a comparative intervention or exposure in the cited administrative or program source. No comparator is reported for this source-bounded claim.
Outcome or endpoint
The program, policy, regulatory, standards, or development state stated in the claim supported; no clinical outcome is inferred.
Follow-up duration
No clinical follow-up duration is reported for this source-bounded claim.
Effect estimate and uncertainty
No clinical effect estimate is reported. No statistical uncertainty estimate is reported; uncertainty is bounded through the source scope and stated limitations.
Harms/adverse events
No harms or adverse-event analysis is reported for this source-bounded administrative or program claim.
Missing data and attrition
Not reported; this record does not analyze participant-level data. Not applicable or not reported; no longitudinal participant cohort is analyzed.
Bias/limitations
The request's comment period ended December 1, 2025. It is discussion-only—not draft or final guidance, an adopted framework, an enforceable requirement, or a policy change.
Applicability/generalizability
Applicable to decisions about the stated program, policy, standard, or development context; not evidence of clinical benefit beyond the cited scope.
Funding/conflicts
Funding or sponsorship is recorded only where the cited source and claim state it; otherwise it is not assessed in this record. No conflict-of-interest assessment is reported for this administrative or program-source record.
Replication or corroboration
The administrative or program claim is anchored to the cited authoritative source; independent clinical replication is not implied.
Decision use
Use as source-bounded context for Clinicians, Innovators, Health-system leaders, Policymakers, Public/patients; do not use as stand-alone evidence of safety, efficacy, coverage, endorsement, or scaled adoption.
Last reviewed / review due
Last reviewed 2026-09-03; review due On material source change; preserve as a dated record.

For this decision

Clinicians · Innovators · Health-system leaders · Policymakers · Public/patients

Primary sources

Related intelligence