The increasing risks posed by climate change, natural disasters, biodiversity loss or pandemics require better planning of how human communities can adapt to their impacts in a sustainable way. Imagining tomorrow’s world and making it visible today so that these communities can understand its complexity and debate its evolution: this is the objective of integrated modeling approaches such as agent-based modeling, which make it possible to explore possible trajectories and identify bifurcations, particularly those related to emerging risks and the vulnerability of populations.

These approaches raise crucial challenges in computer science, AI and software engineering, as they rely on the construction and simulation of large-scale artificial worlds in which the behaviors of human societies and their environment must be represented in detail. Yet, they are becoming progressively essential to explore, and even build, scenarios that respond to the sometimes contradictory constraints of the various actors.

Because modeling offers freedoms that practice alone does not allow, and because it can bring out new representations capable of accompanying, or even stimulating, indispensable ruptures, it can help us to reexamine our practices and thus accompany the transition to more sustainable lifestyles. Current research perspectives, illustrated by examples drawn from the work of IRD researchers with Vietnamese partners, will be presented, with the main objective of opening a fruitful discussion on the role of models, and in particular agent-based models, in the new emerging discipline called “Sustainability Science”.