The FDA and EMA have jointly issued Good AI Practice principles for drug development, outlining shared expectations for responsible AI use, governance, and patient safety as artificial intelligence becomes more widely used across the pharmaceutical life cycle
Written By: Pharmacally Medical News Desk
FDA and EMA Release Joint Good AI Practice Principles for Drug Development On 14 January 2026, the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) jointly released a new document outlining Good Artificial Intelligence (AI) Practice in Drug Development. The guidance reflects a shared effort by both regulators to align expectations as AI becomes more widely used across the pharmaceutical development process. The agencies emphasize that while AI can speed up development and improve analysis, it does not change the basic rules for approving medicines drugs must still meet established standards for quality, safety, and effectiveness, with benefits clearly outweighing risks.
Growing Use of AI
AI now supports many stages of the drug life cycle, including laboratory research, clinical trials, manufacturing, and post-marketing safety monitoring. These systems are complex and evolve over time, raising challenges in reliability, control, and oversight. When integrated properly with human judgment, AI can shorten timelines, enhance safety monitoring, and predict drug performance in humans, potentially reducing animal testing reliance.
Core 10 Principles
The guidance outlines 10 shared principles for responsible AI use. These form a risk-based framework prioritizing patient safety.
- Human-centric by design: Prioritizes patient safety, ethical use, and human oversight in all AI applications.
- Fit for purpose: Matches AI capabilities to specific roles, clearly defining uses, limitations, and decision-making influence.
- Risk-based: Scales validation, oversight, and controls to the system’s risk level and patient impact.
- Multi-disciplinary expertise: Involves diverse teams for development, deployment, and monitoring.
- Data governance: Ensures suitable data quality, documentation, and transparency in sources/processing.
- Sound design practices: Follows established software/system engineering aligned with GxP requirements.
- Performance evaluation: Conducts ongoing assessments, real-world monitoring, and periodic reviews.
- Transparency and documentation: Provides clear, accessible information on AI function, data, performance, and limits.
- Postmarket monitoring: Continuously tracks safety, effectiveness, and changes throughout the lifecycle.
- Stakeholder communication: Shares understandable details with regulators, users, and patients as relevant.
Governance and Oversight
AI development requires sound practices, appropriate data, and full documentation of steps for verification under GxP standards. Performance monitoring is continuous, addressing issues like data shifts or accuracy drift. Clear communication in plain language builds trust across stakeholders.
AI in Action
We have seen recent collaborations that illustrate these AI applications in practice. Nvidia and Eli Lilly announced a $1B, five-year AI co-innovation lab in San Francisco’s Bay Area in early 2026, while Roche advances brain drug research tools with AI specialists, and Owkin partners with Amgen, AstraZeneca, and Sanofi to predict clinical outcomes, discover biomarkers, and optimize trial design and patient stratification.
Global Alignment Ahead
This document starts global harmonization, evolving with AI advances to balance innovation and patient protection. Sponsors must detail AI roles/limitations in submissions; check FDA guidance or EMA page. For pharmacally.com readers, emphasize GxP-aligned documentation to meet these expectations.
References
Guiding principles of good AI practice in drug development, January 2026, https://www.ema.europa.eu/en/documents/other/guiding-principles-good-ai-practice-drug-development_en.pdf
Guiding Principles of Good AI Practice in Drug Development, US FDA, January 2026, https://www.fda.gov/media/189581/download
Guiding Principles of Good AI Practice in Drug Development, 14 January 2026, https://www.fda.gov/about-fda/artificial-intelligence-drug-development/guiding-principles-good-ai-practice-drug-development
NVIDIA and Eli Lilly Bet $1B on AI-Driven Drug Discovery, 14 January 2026, https://pharmacally.com/nvidia-and-eli-lilly-bet-1b-on-ai-driven-drug-discovery/
AI and machine learning: Revolutionising drug discovery and transforming patient care, 30 January 2026, https://www.roche.com/stories/ai-revolutionising-drug-discovery-and-transforming-patient-care
Sanofi invests $180 million equity in Owkin’s artificial intelligence and federated learning to advance oncology pipeline, 18 November 2026, https://www.sanofi.com/en/media-room/press-releases/2021/2021-11-18-06-30-00-2336966
Owkin announces partnership with AstraZeneca to develop an AI gBRCA pre-screen solution for breast cancer, 02 October 2024, 02 October 2026, https://www.owkin.com/newsfeed/owkin-announces-partnership-with-astrazeneca-to-develop-an-ai-gbrca-pre-screen-solution-for-breast-cancer
Owkin and Amgen use AI to improve cardiovascular risk prediction, 16 December 2021, https://www.owkin.com/newsfeed/owkin-and-amgen-use-ai-to-improve-cardiovascular-risk-prediction