Medicines resulting from research and development (R&D) activities include medicinal products developed through a multi-stage process of preclinical and clinical trials conducted to demonstrate their quality, safety, and efficacy. This process is lengthy, expensive, and fraught with a high risk of failure. It is financed by both private and public funds (according to research by Claudie Wild, Ozren Sehic, Louise Schmidt, and Daniel Fabian, “Public contributions to R&D of medical innovations: A framework for analysis,” 26 publications were identified, which found that half of all approved medicines and >90% of target medicines are linked to public sector institutions or their funding), including EU funds and national innovation support programs. From a pharmaceutical law perspective, R&D medicines are subject to specific regulatory requirements at all stages of the product lifecycle, from clinical trials to marketing authorization and pharmacovigilance. In this context, the proposed changes to EU pharmaceutical regulations, concerning data standardization and information transparency, constitute part of a broader regulatory framework for the functioning of the innovative medicines market. The research and development process is as follows:
In recent years, the importance of artificial intelligence (AI) in drug development, evaluation, and monitoring has grown significantly. AI technologies have the potential to accelerate research, improve predictions of drug efficacy and safety, and reduce the need for animal testing. At the same time, their use presents new challenges. AI models can make errors, be susceptible to unforeseen risks, or use data in a non-transparent manner. To fully realize the benefits of AI while minimizing risks, it is essential to establish clear and common principles for the use of these technologies. In response to these challenges, the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) have jointly developed ten principles of good practice for the use of AI in the drug lifecycle. This document is fundamental and a framework, not a binding legal regulation – it provides general directions and guidelines that should guide drug manufacturers, applicants, and regulators. These principles indicate how AI should be designed and used to ensure it is ethical, safe, transparent, and based on reliable data. The ten principles also identify areas where international regulators, standards-setting organizations, and other collaborating entities can work together to promote good practice in drug development. These areas of collaboration include: conducting scientific research, creating educational tools and resources for market participants, international harmonization, and developing consensus standards. To facilitate initial analysis, these principles can be grouped into three logical pillars. Principles 1-3 address organizational foundations and people, focusing on interdisciplinary team expertise and ensuring that AI remains under human control within specific, established governance processes. Principles 4-7 address technical quality and model integrity, addressing the “heart” of the technology. Principles 8-10 address accountability and lifecycle, defining standards for documentation, clear communication with users, and continuous monitoring of the model after its implementation. Below is a detailed summary of the 10 principles of good practice for AI in the drug lifecycle:
