AI*IA 2017 Advances in Artificial Intelligence

In this work we present a platform shaping citizens’ behavior in case of critical hydrogeological phenomena that can be manipulated in order to realize many possible scenarios. Here the citizens (modeled through cognitive agents) need to identify the risk of a possible critical events, relying of their information sources and of the trustworthiness attributed to them. Thanks to a training phase, the agents will be able to make a rational use of their different information sources: (a) their own evaluation about what could happen in the near future; (b) the information communicated by an authority; (c) the crowd behavior, as an evidence for evaluating the level of danger of the coming hydrogeological event. These weather forecasts are essential for the agents to deal with different meteorological events requiring adequate behaviors. In particular we consider that the authority can be more or less trustworthy and more or less able to deliver its own forecasts to the agents: due to the nature itself of the problem, these two parameters are correlated with each other. The main results of this work are: (1) it is necessary to optimize together both the authority communicativeness and trustworthiness, as optimizing just one aspect will not lead to the best solution; (2) once the authority can reach much of the population it is better to focus on its trustworthiness, since trying to give the information to a larger population could have no effect at all or even a negative effect; (3) the social source is essential to compensate the lack of information that some agents have.