EU and US Regulators Reach Landmark Accord on AI Principles in Drug Development
The European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) have established new AI principles in drug development, aiming to reduce regulatory divergence between the major markets of the European Union and the United States.
This landmark accord is welcomed by industry associations, as it strengthens harmonization across regions; however, they emphasize the need for more concrete steps moving forward.
The Shift from Monitoring to Principles-Based Regulations
As AI technologies become increasingly integrated into evidence generation and analysis in drug development, regulators are transitioning from mere monitoring to establishing principles-based guardrails. This shift aims to enhance accountability, integrity, and performance of the emerging technology.
The accord is expected to significantly influence global AI use in drug development, as the regulatory weight of EMA and FDA decisions sets standards worldwide.
Guiding Principles and Renewed Cooperation
The AI principles, published on January 14, mark the culmination of a two-year process aimed at addressing the regulatory divergence that has posed a significant barrier to digital innovation in the pharmaceutical sector. According to Olivér Várhelyi, European Commissioner for Health and Animal Welfare, these guiding principles are a first step towards renewed EU-US cooperation in the field of novel medical technologies.
The goal is to preserve a leading role in the global innovation race while ensuring the highest level of patient safety.
Industry Response
Industry associations, such as the European Federation of Pharmaceutical Industries and Associations (EFPIA), view the joint effort of EMA and FDA in developing these principles as an important move towards global regulatory convergence. EFPIA represents the interests of 36 national associations and 40 pharmaceutical companies operating in Europe.
Comprehensive Governance of AI in Drug Development
The AI principles are designed to govern the use of AI technology throughout its lifecycle in drug development—from early-stage drug discovery to clinical trials and post-market safety monitoring. This ensures that patient safety and ethical integrity are prioritized, emphasizing a human-centric approach and oversight.
With these new principles, EMA and FDA aim to dismantle the “AI black box,” which often leaves users and patients unaware of the processes behind result generation. Regulatory bodies state that AI must be applied in well-defined contexts and be understandable to all parties involved.
Addressing Challenges and Enhancing Accountability
A crucial aspect of the joint document is its focus on tackling the issue known as “shadow use,” where analysts work with large language models (LLMs) in daily tasks while leadership may overlook this practice. Enhanced oversight is required to counteract this.
Furthermore, the principles mandate that data scientists collaborate closely with clinical leads throughout the drug development process, ensuring that clinical teams understand the technical tools being utilized.
Instead of relying on single validation, continuous monitoring for “data drift”—the degradation of AI performance over time due to changes in the underlying data environment—is now a requirement.
Next Steps and Future Implications
While the new AI principles align with existing regulations, they do not drastically alter how AI is currently employed in the European industry. However, they lay the groundwork for a unified language in medical technology development, which could have extensive implications for global AI regulation in drug development.
According to EFPIA, these principles facilitate a more coherent environment for scaling AI tools globally and for engaging with regulators consistently. While the guidelines represent a solid foundation for minimizing duplicative or divergent requirements across regions, they remain high-level. Further steps toward establishing shared terminology, definitions, and concepts are anticipated in the near future.
