AI4AOP: Current Thinking

The rapid expansion of mechanistic biological data, together with regulatory pressure to reduce animal testing, is driving interest in how artificial intelligence (AI) can support New Approach Methodologies (NAMs) and regulatory toxicology. Within this context, the Adverse Outcome Pathway (AOP) framework provides a structured and internationally endorsed basis for linking mechanistic evidence to regulatory-relevant adverse outcomes. This slide deck presents current thinking on how AI can assist AOP development, use, and community practices, while preserving scientific transparency and regulatory trust. It introduces the AI4AOP initiative, with a focus on two ongoing activities: the use of AI to support ontology annotation and harmonisation of AOP Key Events, and a Reverse Mentoring exercise in which early-career, AI-savvy participants and experienced AOP experts jointly explore effective and responsible AI use. The slides are based on a presentation delivered at the Global Summit on Regulatory Science (GSRS25) in September 2025 and will be updated over time to reflect evolving insights and practical experience. Last update: 2026-01-27.