Offline and online adaptation in prosocial games

Personalization and maintenance of high levels of engagement still remain two of the main challenges in the design of serious games. Towards this end, in this paper we propose a novel adaptation approach for both online and offline adaptation in prosocial games. In this paper, we describe the implementation of an artificial intelligence driven adaptation manager, whose purpose is to direct players towards game content the players are most likely to enjoy (measured in their engagement responses). As a consequence, we demonstrate how the adaptation manager can be used to increase the chances of players attaining the game's specific prosocial learning objectives. Each mechanism (offline and online) processes different information about the player and concerns different types of factors affecting engagement and prosocial behavior. More specifically, the online mechanism maintains a player engagement profile for game elements related to the provision of Corrective Feedback and Positive Reinforcement, in order to adapt existing game content in real time. On the other hand, off-line adaptation matches players to game scenarios according to the players' prosocial ability and the game scenarios' ranking. The efficiency of the proposed adaptation manger as a tool for enhancing students' prosocial skills development is demonstrated through a small scale experiment, under real-conditions in a school environment, using the prosocial game of Path of Trust.