Marine environments undergo rapid changes under the influence of various pressures (human footprint, climate change). Design indicators to monitor the status of these ecosystems and their biodiversity is critical. However, such a monitoring is based on data, and data is a real challenge in environmental science: sparse in space and time, heterogeneous, highly unbalanced, inconsistent because collected according to different protocols subject to multiple biases, or difficult to interpolate/extrapolate because resulting from complex and dynamic processes, these features are all obstacles to the creation of indicators.

In recent years, Artificial intelligence has led to significant advances in many fields, and could be a way of overcoming these difficulties.