Abstract

Advances in artificial intelligence (AI) are creating new opportunities for biodiversity monitoring by enabling the integration of diverse data sources, including species observations, environmental measurements, remote sensing imagery, and citizen science records. Yet, translating these diverse data streams into ecological insights requires addressing fundamental statistical challenges and fostering interdisciplinary collaboration among statisticians, ecologists, and AI researchers.

This talk will explore the role of statistical methodology in AI-driven biodiversity monitoring, with particular emphasis on ecological niche modelling. These models provide a framework for understanding species–environment relationships and predicting species distributions under changing environmental conditions. Integrating AI techniques offers the potential to capture complex nonlinear relationships, leverage high-dimensional data, and improve predictive performance. However, sampling bias, data heterogeneity, imperfect detection, spatial dependence, and uncertainty quantification raise important questions about model reliability, interpretability, and methodological extensions. These issues create significant opportunities for statistical research in developing robust, transparent, and uncertainty-aware AI methods for biodiversity monitoring.