Towards a social Artificial Intelligence

Artificial Intelligence can both empower individuals to face complex societal challenges and exacerbate problems and vulnerabilities, such as bias, inequalities, and polarization. For scientists, an open challenge is how to shape and regulate human-centered Artificial Intelligence ecosystems that help mitigate harms and foster beneficial outcomes oriented at the social good. In this tutorial, we discuss such an issue from two sides. First, we explore the network effects of Artificial Intelligence and their impact on society by investigating its role in social media, mobility, and economic scenarios. We further provide different strategies that can be used to model, characterize and mitigate the network effects of particular Artificial Intelligence driven individual behavior. Secondly, we promote the use of behavioral models as an addition to the data-based approach to get a further grip on emerging phenomena in society that depend on physical events for which no data are readily available. An example of this is tracking extremist behavior in order to prevent violent events. In the end, we illustrate some case studies in-depth and provide the appropriate tools to get familiar with these concepts.