When does artificial intelligence replace process-based models in ecological modelling?

•Artificial neural networks (ANN) are competitors of process-based models (PBM).•Sven Jørgensen once wrote that ANNs cannot replace some PBMs.•There is still no sign that ANNs are replacing any PBMs.•There are methods for solving differential equations by using ANNs.•Converting PBMs into an ANNs enhances predictions based on the big data. Sixteen years ago, Sven Jørgensen, a founder of Ecological Modelling, wrote that artificial neural networks could be very useful in most cases but cannot replace biogeochemical models based on conservation principles. The literature study shows no sign that artificial neural networks are replacing any process-based models in ecological modelling, although the recent efforts in developing the methods for solving differential equations by using neural networks enable turning the process-based models into neural networks without loss of their theoretical rigor. It seems that a well-orchestrated research program is needed to promote turning existing process-based models into neural networks aimed at taking advantage of opportunities offered by the big data revolution.