Jan 22, 2026
3 min read
The industry agrees that AI holds the potential to transform health, as long as progress is made within clear regulatory frameworks.
The pharmaceutical industry has drawn attention to the Ten Commandments of Best Practices for Using Artificial Intelligence (AI) in Drug Development, published jointly by the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA).Both organizations recognize that this technology will accelerate biomedical innovation and seek to build a common foundation for safe adoption.
A document published by the EMA and the FDA sets out ten principles that guide the prudent use of artificial intelligence, understood as a set of systems capable of generating or evaluating scientific evidence in all phases of a drug's life cycle.From pre-clinical research to marketing and post-marketing, organizations find these tools can provide greater efficiency and sophistication.
The industry recognizes that artificial intelligence opens a window of opportunity to transform health, if progress is made within a clear legal framework that aligns with the digital reality.The sector, which is already at the forefront of implementing these technologies, emphasizes the importance of implementing ethical standards and ensuring that innovation benefits society and patients.
To speed up AI regulation and reduce animal testing
Regulators emphasize that the use of AI can help reduce the time it takes for new drugs to reach the market, strengthen regulatory benefits and improve pharmacovigilance.Among the benefits mentioned is also the possibility of reducing dependence on animal models in the early stages of research thanks to more accurate predictions of efficacy and toxicity.
With this initiative, the EMA and the FDA seek to facilitate a common working environment between regulators, industry and researchers, with the aim of making artificial intelligence a reliable tool for accelerating biomedical innovation without compromising the quality, safety or effectiveness of medicines.For the industry, this is a critical step towards a more technologically advanced and, above all, patient-centric future of healthcare.
This document establishes 10 principles that can be used by the various organizations involved, from drug manufacturers to applicants or marketing authorization holders.It also highlights the importance of strong international partnerships to promote responsible innovation.
- Human-centered design
The development and use of AI technologies must be consistent with human-centered ethical values.
- Risk-based approach
The level of validation and oversight of AI systems should be appropriate to the risks they may pose to the specific model and context of use.
- Monitoring of standards
AI technologies must comply with relevant legal, ethical, technical, scientific, cybersecurity and regulatory standards, including Good Practice (GxP).
- Clear user experience
The context of use in AI systems, their function and scope must be well established.
- Multidisciplinary approach
The development of all artificial intelligence systems should be approached in a multifaceted manner.
- Administration and documentation
The origin of data sources, their treatment and analysis decisions should be documented in a detailed, traceable and verifiable manner in accordance with good practice.Thus, proper control as well as data security and confidentiality will be maintained throughout the entire life cycle of the technology.
- Design and development of models
The development of artificial intelligence technology must follow best practices in model design, systems, and software engineering to leverage data with interpretability, understandability, and predictive performance in mind.Good model and system development will promote transparency, reliability, generalizability and robustness of AI technology, contributing to patient safety.
- Risk-based performance evaluation
Risk-based performance evaluation analyzes the entire system, including human-AI interactions, using data and metrics appropriate to the intended use case and supported by verification of predicted performance using properly designed test and evaluation methods.
- Life cycle management
Risk-based quality management systems are implemented throughout the lifecycle of AI technologies, including support for discovery, evaluation, scheduled monitoring, and periodic re-evaluation of data.
- Clear and relevant information
Plain language should be used to provide clear, accessible and informative information to the audience, including users and patients, to make the context of use, performance, limitations, underlying data, updates and AI technology understandable and comprehensible.
