Abstract.Games have a great potential as learning tools, in particular, because they provide means for players to safely explore and fail, and because they promote personal emotional experiences. To be successful, games must present a good coverage and fidelity of the interaction experience regarding the target learning goals. In the case of learning of social skills, which is one prominent area of application of games, the use of AI characters with socioemotional agency has great potential value. These characters may increase the range of social situations that players can explore (coverage). However, in order to achieve that the AI characters need to be able to present good social behaviour (fidelity). Although, several examples of computational models to achieve this can be found, developing these models remains a challenging research question.
Applied Games and LearningGames have been used for other purposes beyond entertainment for several years. For example, as tools to engage people in crowdsourcing tasks, e.g. The use of games for learning is sustained in the strong connection between play and learning, noted in child development studies [7], an idea that was already discussed in the seminal work of John Locke and Jean-Jacques Rousseau in the XVII and XVIII centuries. In turn, learning has been regarded as an important activity in games for promoting fun. Following the Natural Funativity theory [8], which presents fun as a reward for learning, a game that has nothing to teach to a player will become boring and will, therefore, be abandoned.The work of Clark Abt [9], in 1970, fostered the idea of having games directly presented as tools for learning. He coined the term serious games as games that have "an explicit and carefully thought-out educational purpose and are not intended to be played primarily for amusement". Abt argued for the development of new games taking learning goals as their main concern. Since then, serious games have grown as a research field and the term has evolved (commonly, referred now as applied games or games with purpose) to incorporate other serious applications besides education. Education and training are still, nevertheless, regarded as one of the application fields with most potential.
Why Applied Games Work for LearningThere are two main reasons why (good) applied games work as learning tools. First, because they enable practice, or, more importantly, because they support proactive exploration and failure. While playing, learners are motivated to explore the range of options presented in the game in the search for specific outcomes. Intended outcomes are not easily achieved, which means that players fail in the process until they achieve a good outcome. Games are good at supporting players recovering from those failures, because they do not imply major consequences in real-life. This means that players can safely fail. They do not fear failure too much and are, therefore, encouraged to explore the different options that the game offers and, for that reason, gain understand...