Regulators across the U.S., EU, and UK are rapidly building new frameworks to manage the growing use of Artificial Intelligence in drug development, where opaque and adaptive models analyse large data sets and increasingly influence regulatory and clinical decisions. The risk posed by an Artificial Intelligence system depends on context and impact on decision-making, and failures such as hallucinations or embedded bias could have severe consequences for patient safety. Authorities are therefore pursuing tools and principles that safeguard public health and research integrity while still promoting adoption of Artificial Intelligence throughout the medicines lifecycle.
On 14 January 2026, the European Medicines Agency and the U.S. Food and Drug Administration agreed to common principles for the use of Artificial Intelligence in drug development, including alignment with human-centric values, robust data governance and risk-based performance assessments of whole systems. In parallel, international regulators are harmonising policies for Artificial Intelligence and machine learning enabled medical devices used in settings such as decentralised clinical trials, where wearables, apps and telemedicine generate real-world evidence. Collaborative work between the U.S. Food and Drug Administration, the UK Medicines and Healthcare products Regulatory Agency and Health Canada has produced Good Machine Learning Practice principles in October 2021, guiding principles for Predetermined Change Control Plans in October 2023 and transparency principles in June 2024, all aimed at ensuring safe, effective, and clearly explained Artificial Intelligence medical technologies that can be updated without repeated regulatory submissions while maintaining oversight.
In the U.S., Artificial Intelligence or machine learning enabled software that meets the definition of a medical device remains under the U.S. Food and Drug Administration’s authority, and the agency is using non binding guidance, action plans and lifecycle management recommendations to adapt its traditional model to adaptive technologies. At the Consumer Electronics Show on 6 January 2026, Commissioner Dr. Martin Makary said the agency is developing a “smarter” framework that moves in a “deregulatory direction for low-risk products” and relies more on post market monitoring while planning further Artificial Intelligence specific guidance. In the EU, the forthcoming Pharma Package agreed on 11 December 2025 is expected to enter into force in 2026 with a 2 year transition, with reforms intended to support broader use of Artificial Intelligence in regulatory decision-making, while the proposed Biotech Act of 16 December 2025 would mandate the European Medicines Agency to issue guidance and create trusted testing environments and data quality accelerators for Artificial Intelligence enabled biotechnology, with final adoption unlikely before late 2026. Regulation (EU) 2024/1689, the Artificial Intelligence Act, introduces tiered risk classes, and although most Artificial Intelligence systems used in drug development are not high risk, they will face enhanced obligations where they intersect with medical device rules, prompting calls to simplify overlapping requirements.
In the UK, the Medicines and Healthcare products Regulatory Agency created the non statutory National Commission into the Regulation of Artificial Intelligence in Healthcare on 26 September 2025 to review existing rules and design a new health care Artificial Intelligence framework for publication in 2026, supported by a broad Call for Evidence to capture views from domestic and international stakeholders. Across jurisdictions, these initiatives share a focus on safe and responsible use of Artificial Intelligence, aiming for an international regulatory environment that speeds the path from discovery to approval while preserving rigorous standards for safety, transparency and accountability. Success will depend on aligning new guidance, commissions and principles with established and emerging legislative instruments so that innovation in Artificial Intelligence driven drug development and biotechnology does not outpace the protections designed to govern it.
