Faculty: Carolyn Troiano
This webinar addresses how pharmaceutical and biologics developers establish AI/ML model credibility for regulatory decision-making, connecting context of use, model risk, data relevance, performance, uncertainty, and supporting evidence to determine whether model outputs are sufficiently reliable for their intended regulatory purpose.
AI/ML models are increasingly being considered to support decisions across drug development, including nonclinical assessments, clinical development, and regulatory evidence generation. However, strong predictive performance alone does not establish that a model is credible for a particular regulatory purpose. FDA's risk-based credibility framework emphasizes the model's context of use and the consequences of an incorrect decision, making the required evidence dependent on how the model will influence regulatory conclusions.
A model may perform well during development yet rely on unrepresentative datasets, poorly characterized assumptions, or validation conditions that differ from its intended application. These weaknesses become consequential when model outputs influence regulatory submissions or development decisions. This webinar addresses how to define credibility requirements, evaluate data and model limitations, establish fit-for-purpose performance evidence, and document the scientific rationale supporting regulatory reliance on AI/ML outputs.
The central consideration is whether the available evidence is sufficient for the model's defined context of use and associated decision risk. The session addresses data relevance, model applicability, uncertainty, performance limitations, and the justification needed when AI/ML outputs contribute to consequential regulatory decisions.
The session follows FDA's proposed credibility assessment framework, from defining the regulatory question and model context of use through risk assessment, credibility planning, evidence generation, and documentation. Practical scenarios illustrate how deficiencies in datasets, validation approaches, or model assumptions affect the acceptability of supporting evidence.

Carolyn Troiano has more than 45 years of experience in computer system validation in the pharmaceutical, medical device, biotechnology, tobacco, and other FDA-regulated industries. She is currently an independent consultant, advising companies on FDA compliance, Computer System Validation (CSV), and large-scale IT system implementation projects. Carolyn participated in the FDA/Industry Partnership to develop 21 CFR Part 11, the FDA’s Guidance for Electronic Records and Electronic Signatures. During her career she has provided consulting and training, including CSV, 21 CFR Part 11, Data Integrity, and many other related compliance topics of interest to the life science industries.