Pharmaceutical regulators in Europe, the UK, the US and Brazil are moving quickly to define how artificial intelligence will be used across the drug lifecycle, creating a complex and fast-changing environment for companies. In Europe, the controversial European Union Artificial Intelligence Act could impose restrictive rules on pharmaceutical research and development, and there is growing concern that overregulation could slow innovation. Legal experts argue that if the European Union Artificial Intelligence Act is applied in a way that tightly constrains research tools and data use, the UK could position itself as a more attractive base for pharmaceutical Artificial Intelligence development by offering a more flexible regulatory framework. Parallel to this, trade discussions between India and the European Union are exploring lower tariffs on pharmaceuticals and a separate investment protection agreement that could affect data exclusivity, underscoring how regulation, trade and data policy are becoming intertwined.
Drug discovery and development are emerging as early test beds for Artificial Intelligence governance. Lawyers warn that although using artificial intelligence in drug discovery and development could have huge benefits for companies, a critical challenge is determining how proprietary and real world data sets are used to train models and how intellectual property will be protected. Regulators are starting to respond by laying out high level expectations. The US Food and Drug Administration and the European Medicines Agency have jointly issued ten guiding principles for pharmaceutical companies using Artificial Intelligence across the drug lifecycle, emphasizing human centric design, a risk based approach and lifecycle management. In the UK, the Medicines and Healthcare products Regulatory Agency is inviting industry to help shape bespoke rules for Artificial Intelligence in health care, while also becoming the first major regulator to require sponsors to disclose Artificial Intelligence and machine learning use in good clinical practice pre inspection dossiers for clinical trials.
Regulators are also integrating Artificial Intelligence into their own operations and encouraging new methodologies. The US Food and Drug Administration reports “scores” of active investigational new drug applications using Artificial Intelligence in cell and gene therapies, while warning that a lack of good quality training datasets still limits decision support tools. The agency is rolling out agentic Artificial Intelligence tools following the earlier launch of its generative system Elsa, and is offering a “road map” for using Artificial Intelligence to make better use of small datasets in rare diseases. The UK is tying Artificial Intelligence to a broader shift away from animal testing by targeting specific studies such as contamination tests and pharmacokinetic analyses for replacement with organ on a chip systems and computational models. Across research and regulatory affairs, case studies range from attempts to use generative tools to decode complex European Union pharmaceutical reforms, which have exposed current limitations, to efforts in Parkinson’s disease cell therapy and predictive safety analytics, where companies and regulators see Artificial Intelligence as a route to more efficient and reliable development pipelines.
