Artificial Intelligence for ecosystem services science: The AI4ESS framework from a community perspective

Ecosystem services (ES) research draws on heterogeneous biophysical, socio-economic, and valuation data that are often collected at different scales and structured using incompatible classifications. This paper presents a community-based synthesis of how Artificial Intelligence (AI) can be responsibly embedded in Ecosystem Services Science (ESS) through the AI4ESS framework, developed from two structured expert dialogues organized under the Ecosystem Services Partnership (ESP) network. Drawing on interdisciplinary insights from researchers, practitioners, and policymakers, the AI4ESS framework is defined as a four-dimensional conceptual structure integrating challenges, ethics, data and models, and opportunities to guide transparent and equitable AI adoption in ES research and in subsequent decision support. The AI4ESS framework extends approaches such as GeoAI and AIXES by integrating ethical, governance, and knowledge-inclusion principles alongside technical considerations. It identifies key priorities for advancing responsible AI in ESS: improving data transparency, promoting explainable and open models, establishing community-driven ethics guidelines, and fostering collaboration be-tween domain and AI experts. This synthesis provides a conceptual foundation for integrating AI into ES valuation and decision support. By articulating actionable pathways, the framework contributes to the development of trustworthy, inclusive, and scalable AI applications that strengthen sustainability governance and evidence-based ecosystem management.