PharmacovigilanceOfficial

EMA sets out how AI is moving from concept to practice in medicine regulation

🇪🇺 EMA · Wed 11 Nov 2026

A new EMA web page outlines expected outcomes of AI use across the EU network and compiles a timeline of milestones from 2021 to August 2026.

The European Medicines Agency (EMA) has published a web page describing how it and the wider European medicines regulatory network are moving from concept to practice in using artificial intelligence (AI). The page was first published on 11 November 2026. EMA says the aim is concrete outcomes, such as helping develop higher-quality medicines more quickly.

EMA defines AI as computer systems that use data to analyse information and generate outputs such as predictions, recommendations, content or decisions. Some systems can adapt and improve with experience. The agency says this is consistent with the EU AI Act, Regulation (EU) 2024/1689. It adds that the shift in regulation mirrors industry, which uses AI at every stage of the medicine lifecycle, from discovery and development through manufacturing to safety monitoring.

EMA links advances in data analytics, automation, machine learning and digital twins to opportunities in product quality, manufacturing efficiency, supply chain resilience and regulatory decision-making. It expects the transition to speed up medicine development and to optimise clinical evidence and quality, including clarity on whether, and in which populations, a medicine works and is safe. It also expects faster assessment of applications, more precise advice for industry and a workforce able to benefit from AI.

The page includes a non-exhaustive timeline. Recent entries include a scientific article on regulatory research priorities for AI, co-signed by EMA staff, in Clinical Pharmacology & Therapeutics (August 2026); the EMA AI Observatory report for 2025 (June 2026); and the rollout of Microsoft Copilot 365 to EMA staff, subject to licensing training, managerial approval and monitoring of use (June 2026). Others are updated FAQs on the AI-enabled 'Scientific Explorer' knowledge mining tool (March 2026), the NDSG 2026-2028 workplan on data and AI (March 2026) and common EMA-FDA principles for AI in medicine development (November 2025).

Earlier milestones listed include the first CHMP qualification opinion on an AI-based development methodology (March 2025), the network strategy to 2028 and guiding principles on large language models in regulatory science (both September 2024), and a reflection paper on AI in the medicinal product lifecycle following consultation in 2023. The page itself does not set out new rules or obligations.

What it means for your team

  • Regulatory teams should expect AI to feature in EU network assessment and advice, with EMA citing faster assessments and more precise industry advice as goals.
  • Sponsors using AI in development can consult the reflection paper, the LLM guiding principles and the common EMA-FDA principles listed on the page.
  • The page is a consolidated overview and timeline, not new guidance; no new requirements are announced.
  • The AI Act definition of AI underpins EMA's framing, which is relevant for companies aligning internal terminology.

Official source: EMA ↗

Article written by VigiEye from the official source (EMA). AI-assisted; check the source before acting.

More Pharmacovigilance All →