AI-Driven Digital Twins in the Pharmaceutical Industry: Model-Informed Manufacturing, Quality, and Regulatory Decision-Making

The pharmaceutical industry is increasingly adopting artificial intelligence(AI)-driven modeling frameworks to address rising system complexity, stringent regulatory expectations, and the demand for robust, efficient manufacturing processes. AI-driven digital twins have emerged as a central enabling technology in this context, providing continuously updated virtual representations of pharmaceutical processes and products that integrate mechanistic understanding with data-driven learning. This review examines the role of AI-driven digital twins in model-informed pharmaceutical manufacturing and quality systems, with an emphasis on hybrid mechanistic and machine-learning (ML) architectures, real-time model updating, and predictive process control. It discusses how these systems support quality by design, continuous manufacturing, and real-time release testing by enabling enhanced process understanding, deviation prediction, and lifecycle optimization beyond traditional static models. Particular attention is given to model validation, credibility assessment, uncertainty quantification, and regulatory acceptance, which are critical for the deployment of AI-driven digital twins in regulated pharmaceutical environments. By synthesizing current methodologies, regulatory perspectives, and emerging industrial applications, this article positions AI-driven digital twins as a foundational technology for next-generation pharmaceutical manufacturing and regulatory decision-making.