Regulation — pharma-specific

FDA CDER AI Framework — the Center for Excellence in Regulatory Decision Science stance on AI in drug review

CDER's Center for Excellence in Regulatory Decision Science (CERSI-adjacent) sets the drug-review-side view on AI, complementing the CDRH SaMD Action Plan. For sponsors, this is the stance the drug reviewers — not the device reviewers — bring to Type C meetings and NDA/BLA submissions.

What the framework actually says

"FDA considers AI to be a family of technologies that will play an increasing role in drug development, from discovery through post-market surveillance … The Agency's approach is risk-based and lifecycle-oriented, with expectations for validation, monitoring, and human oversight scaled to the context of use." CDER discussion paper: Using Artificial Intelligence and Machine Learning in the Development of Drug and Biological Products (May 2023, subsequent CDER framework refinements 2024-2026)

"Sponsors should establish and document the context of use for each AI-based tool employed in the development program … and provide a plan for ongoing performance evaluation." CDER AI Framework — context-of-use + ongoing evaluation expectation

What this means in plain English

Three CDER-side stances the sponsor needs a record for:

Context of use. The AI's role must be documented per use — a trial-eligibility classifier is a distinct context of use from an adverse-event narrative generator.
Risk-based validation. Higher-impact contexts (eligibility, dosing recommendation) need proportionately stronger validation + monitoring evidence.
Ongoing evaluation. Not a one-time validation report — a live monitoring plan.

CDER reviewers ask the "context of use + ongoing evaluation" question directly in Type B/C meetings when AI is invoked in trial conduct.

What triggers the exposure in the sample

Trial-eligibility classifier is a high-risk context of use. 10 high-severity drift events + a P3 exclude-lane rate shift from 22% to 75% is the ongoing-evaluation failure mode CDER expects sponsors to catch. Sponsors that arrive at Type C without an independent monitoring record on a high-risk context-of-use AI face avoidable reviewer pushback.

What the $499 Snapshot shows against this framework

Named context of use (trial-eligibility classifier) with model-version pin
Ongoing-evaluation evidence — 90-day audit period with dated drift-event log
Independent-verifier signature — the evaluation is not a self-attest
3 fix-first items scoped to the context-of-use risk profile

See how the determination binds to context-of-use risk →

How does this help me?

Type B/C meeting stance is set by the paper on file. Independent monitoring evidence reframes the reviewer's opening question from "do you monitor this AI?" to "here is the record."

Read: FDA CDER AI Framework -- what it saves in Type B/C meeting stance →

$499 Snapshot. 3 business days.

Independent-verifier determination scoped to your sponsor's AI surface + 3 fix-first items + signed declaration.

Buy $499

Snapshot credit applies to Baseline ($2,500) or Enterprise Attestation ($35-55K) upgrade within 30 days.