AI MEDICAL
The pharma rulebook of AI
The European Medicines Agency and the U.S. Food and Drug Administration have agreed on 10 shared principles for using AI in medicine development.
The guidance comes as AI use across biopharma grows quickly, affecting everything from drug discovery and clinical trials to manufacturing and safety monitoring.
The regulators say these principles will shape future AI guidance in both the US and EU and help set international standards.
They apply to AI systems used at any stage of a medicine’s lifecycle.
At the centre of the framework is the idea that AI should be human-centred, ethically aligned, and built using a risk-based approach that reflects how it will be used.
Systems must meet existing legal, technical, scientific, cybersecurity, and regulatory standards, with strong controls around data quality, privacy, and documentation.
Before being deployed, AI tools should be tested for performance and human-AI interaction.
Once in use, they must be monitored and reviewed regularly to ensure they remain suitable.
The regulators also stress that AI outputs should use clear, plain language that can be understood by professionals and patients alike.
In short:
US and EU regulators are aligning on how AI should be used in drug development
The focus is on risk management, transparency, and ongoing oversight
Biopharma investment in AI continues to accelerate through deals and acquisitions
No black boxes allowed
European Commissioner for Health and Animal Welfare Olivér Várhelyi said the principles mark an early step in renewed EU-US cooperation on new medical technologies, aiming to support innovation while maintaining high patient safety standards.
The pace of AI adoption is reflected in recent industry deals. Iktos has partnered with Servier and Pierre Fabre, while Insilico Medicines is also working with Servier.
GSK has announced partnerships with Noetik and Helix, and AstraZeneca has acquired Modella AI to bring AI development in-house.
If only we could have a rulebook for deepfakes… - MV