‍How can ICH Q9(R1) guide the implementation of AI/ML in GMP? This article explores how uncertainty, importance and complexity can be used to determine the appropriate validation effort for AI systems, based on key discussions from the EMA Annex 22 Workshop.

The EMA Annex 22 workshop on Artificial Intelligence provided an important opportunity for industry to discuss practical approaches for implementing AI in GMP. This article summarizes the key recommendations presented by industry, focusing on Quality Risk Management (ICH Q9(R1)), AI guardrails, and the practical use of Generative AI and Large Language Models in regulated pharmaceutical environments.