EMA Reflection paper on the use of artificial intelligence in the lifecycle of medicines
Summary
What is happening? The EMA has adopted its revised reflection paper on use of artificial intelligence after reviewing and implementing 1342 comments from 66 stakeholders. The reflection paper provides considerations on the use of AI/ML in the lifecycle of medicinal products, including medicinal products development, authorisation, and post-authorisation. Given the rapid development in this field, the aim of this reflection paper is to reflect on the principles that are relevant for regulatory evaluation when these emerging technologies are applied to support safe and effective development, manufacturing and use of medicines.
Key Messages:
- Legal Frameworks: The final version aligns AI/ML use with new EU laws, including the AI Act, GDPR, Cybersecurity Act, and AI Liability Directive
- Ethical AI: Ethical principles and trustworthy AI guidelines are emphasized, promoting transparency, fairness, and human oversight
- Data Integrity: Strengthened guidance on data protection, anonymization, and managing risks with large language models
- Governance: A comprehensive governance approach should cover AI/ML development, deployment, and continuous risk assessment
- Regulatory Interactions: Early engagement with regulators is encouraged for high-risk AI applications in clinical trials and manufacturing
- Model Testing: Clear guidance is provided on avoiding data leakage and ensuring robust model validation with future-use data
Sources
Link to GRIP for draft: https://gileadconnect.sharepoint.comhttps://gileadconnect.sharepoint.com/sites/GRIP/SitePages/View-Alert.aspx#Id%3D2540
Link to revised reflection paper: https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf
Link to note on implementation of comments: https://www.ema.europa.eu/en/documents/comments/implementation-comments-received-draft-reflection-paper-use-artificial-intelligence-medicinal-product-lifecycle-ema-chmp-cvmp-83833-2023_en.pdf