Artificial intelligence and decision support in applied ecology

Abstract Artificial intelligence (AI) is bringing ecological inference closer to decision-making, producing detections, alerts, trends, and prioritisation outputs. Yet, consensus on how to interpret or act on these remains limited. We synthesise AI adoption in applied ecology and identify governance pressures as models become multimodal, transferable, edge-deployable, and increasingly shaped by large language models. We highlight cross-cutting risks where uptake outpaces oversight, including limits in explainability, validation, data sovereignty, environmental costs, cognitive off-loading, and evidence integrity. Using the British Bat Survey as a representative case study, we show how AI can be operationalised through governance principles spanning benchmarking, transparency, auditability, data governance, equity, and sustainability. Policy implications: We propose a roadmap for responsible AI in applied ecology linking evaluation to decision context, clarifying responsibility, and ensuring AI strengthens rather than displaces accountable decision-making. Read the free Plain Language Summary for this article on the journal's blog.