EMA Convenes AI-in-GMP Workshop to Shape Annex 22 Guidance on Adaptive Models
- EMA hosts two-day multistakeholder workshop (30 June–1 July 2026) to gather expert input on responsible AI use in medicines manufacturing, including guardrails for generative AI and large language models in GMP applications.
- Workshop addresses 2025 stakeholder feedback supporting potential use of adaptive and probabilistic models, reversing draft Annex 22 language that prohibited such technologies in critical GMP areas.
- EMA seeks expert consensus on validation paradigms, quality risk management frameworks, human oversight requirements, and guardrail effectiveness to enable risk-based AI deployment in high-impact manufacturing workflows.
Operational Impact
CRO and CDMO manufacturers must track Annex 22 finalization to understand whether adaptive AI and LLM deployment in GMP-critical processes will be permitted under new guardrail and validation standards, affecting manufacturing automation and compliance roadmaps.
Source: EMA What's New RSS