EMA Workshop regarding Annex 22 "Artificial Intelligence"
The current draft of Annex 22, "Artificial Intelligence", proposes excluding generative AI and large language models (LLMs), as well as probabilistic, dynamic and adaptive AI models, from critical GMP applications. Numerous comments submitted during the public consultation challenged this approach. Instead, the aim should be to support innovation and encourage investment by the pharmaceutical industry.
However, the Annex 22 Working Group considers a comprehensive, well-documented, robust and risk-based control strategy necessary if the scope is to be expanded. How such a control strategy might look was discussed at a workshop with AI experts. The discussions were based on presentations by industry associations, reflecting a shared industry perspective on six predefined topics:
- Regulatory frameworks for adaptive/probabilistic AI models in the GMP environment
- Technical reliability and incident response
- Human oversight and accountability
- Validation and lifecycle management
- Strategic risks and compliance boundaries
- Cybersecurity and outsourced activities
The presentations and subsequent discussions are intended to deepen the Working Group’s understanding of current practices, challenges and emerging approaches to managing risks associated with AI systems in the GMP environment. The insights gained are expected to inform the revision of the scope of Annex 22 and support the development of a risk-proportionate and practical regulatory framework. The aim is to ensure both product quality and patient safety while enabling responsible innovation.
Read more about the outcomes of this workshop in our next news update.
Source:
Meet the GMP Compliance Adviser
The GMP Compliance Adviser is the world's largest knowledge portal for quality management in the pharma business.
The demo access is non-binding and ends automatically.
You may also be interested in the following articles:
