From Virtual Teams to Convergent Evidence: Risk-Based Validation for AI-Driven Exploratory Clinical Analysis

About this session

Role-based AI “Virtual Teams” can generate exploratory clinical analyses by coordinating analysis planning, statistical coding, and narrative summaries. Because generative models are non-deterministic, repeated executions can be treated as stochastic samples from a distribution of plausible analytical ideas. Specifically, we will describe how this AI workflow repeats its processes, compares results across multiple runs, and provides the end user with convergence and variability assessments to enable efficient human-in-the-loop validation and accountability.

Speaker

• AG

  Alejandro Gonzalez

  Senior Manager, Digital Innovation & Execution · Genmab AS

  Speaker