Considerations for the Use of Artificial Intelligence To Support Regulatory Decision-Making for Drug and Biological Products; Draft Guidance for Industry; Availability; Comment Request
Published Date: 1/7/2025
Notice
Summary
The FDA is rolling out a draft guide to help drug and biological product makers use artificial intelligence (AI) safely and smartly when making decisions about their products. This affects companies working on medicines and biologics, aiming to boost safety and quality without slowing things down. Comments on this guide are open until April 7, 2025, so industry folks have time to weigh in before the final rules drop.
Free Policy Watch
New rules are filed every week. Most people never see them.
Pick a topic. PRIA watches every federal rule and tells you when one hits your household.
Pick a topic to get started
Analyzed Economic Effects
5 provisions identified: 4 benefits, 1 costs, 0 mixed.
Risk-based AI credibility framework
The FDA proposes a risk-based credibility assessment framework that sponsors should use to establish and document the credibility of AI models for a specific context of use (COU). The framework asks sponsors to plan, gather, organize, and document credibility evidence and to tailor oversight, performance acceptance criteria, and risk mitigation to the model risk and COU.
Guidance applies across product lifecycle
The draft guidance covers AI models used to produce information or data that support regulatory decision-making in the nonclinical, clinical, postmarketing, and manufacturing phases of the drug and biological product life cycle. If your AI use produces data or information for safety, effectiveness, or quality decisions in those phases, it falls within this guidance's scope.
Certain AI uses excluded from scope
The guidance does not apply to AI used in drug discovery or to AI used solely for operational efficiencies (for example internal workflows or drafting submissions) when those uses do not impact patient safety, drug quality, or study result reliability. Sponsors with those AI uses are not covered by this draft guidance for regulatory decision-making purposes.
Options for early FDA engagement
The draft guidance describes options for sponsors to engage with FDA early to set expectations about appropriate credibility assessment activities based on model risk and COU and to identify challenges. It also notes sponsors may choose to complete the guidance's steps without early engagement and keep postmarketing documentation in standard operating procedures for agency review upon request.
Comment deadline and feedback priorities
FDA is accepting public comments on this draft guidance and the proposed collection of information through April 7, 2025. FDA specifically requests feedback on how well the proposed risk-based credibility framework aligns with industry experience and whether additional guidance on AI in postmarketing pharmacovigilance would be helpful.
Your PRIA Score
Personalized for You
How does this regulation affect your finances?
Sign up for a PRIA Policy Scan to see your personalized alignment score for this federal register document and every other regulation we track. We analyze your financial profile against policy provisions to show you exactly what matters to your wallet.
Key Dates
Take It Personal
Get Your Personalized Policy View
Take the PRIA Score to see how policy affects your household, then upgrade to PRIA Full Coverage for year-round monitoring.
Already have an account? Sign in