EMA and FDA Publish 10 Shared Principles for AI in Drug Development

The European Medicines Agency (EMA) and the U.S. Food and Drug Administration on January 14 jointly published ten guiding principles for good artificial intelligence practice in the development of medicines, EMA announced.

The principles give broad guidance on AI used for generating and monitoring evidence in every phase of a medicine's life, from early research and clinical trials to manufacturing and safety monitoring, the agency said. They are aimed at medicine developers and at applicants for and holders of marketing authorisations.

What the principles say

The joint document, dated January 2026, defines AI as system-level technologies used to generate or analyse evidence across the drug product life cycle, including nonclinical, clinical, post-marketing and manufacturing phases. It applies to drugs and biological products as defined in the United States and medicinal products as defined in the European Union.

The ten principles are: human-centric design; a risk-based approach, with validation and oversight proportionate to the context of use and model risk; adherence to legal, ethical, technical, cybersecurity and regulatory standards, including Good Practices (GxP); a clear context of use; multidisciplinary expertise; data governance and documentation; good model design and development practices; risk-based performance assessment; life cycle management; and clear, essential information.

On data, the document says provenance, processing steps and analytical decisions should be documented in a detailed, traceable and verifiable way, in line with GxP requirements. Performance assessments should evaluate the complete system, including how people interact with the AI, and life cycle management should include scheduled monitoring and periodic re-evaluation, for example to address data drift. Information about an AI tool's context of use and performance should be presented in plain language to its intended audience, “including users and patients,” the principles say.

The agencies write that AI technologies are expected to help reduce time to market, strengthen pharmacovigilance and decrease reliance on animal testing “by improving the prediction of toxicity and efficacy in humans.”

What comes next

EMA said the principles will underpin future AI guidance in the different jurisdictions and support collaboration among regulators and standards bodies. Guideline development in the EU is already under way, building on the AI reflection paper EMA published in 2024, and the principles will be supplemented over time by further EU guidance, according to the agency. The work follows an FDA-EU bilateral meeting in April 2024.

“The guiding principles of good AI practice in drug development are a first step of a renewed EU-US cooperation in the field of novel medical technologies,” said Olivér Várhelyi, the European Commissioner for Health and Animal Welfare, in EMA's announcement.

As Regulatory Focus reported, the document resembles the good machine learning practice principles for medical devices adopted in recent years by regulators in the US, UK and Canada and by the International Medical Device Regulators Forum. Joe Franklin, special counsel at Covington & Burling, told the publication that the document “affirms both the anticipated benefits of AI in drug development as well as many of the AI deployment considerations that companies are building into AI governance.”