Teaching Artificial Intelligence: Potentialities and Limitations of Different Approaches

Abstract

The field of Artificial Intelligence has always had fluid boundaries and a multidisciplinary character, bringing together diverse domains and asking for different perspectives. In light of these characteristics, we examine here a few strategies that can be adopted in introductory Artificial Intelligence classes. We analyze these possible strategies by considering several aspects (theoretical and conceptual nature, historical focus, practical emphasis, and prominence to modern themes) and discuss the potential and limitations of different approaches positioned on a conceptual map. We use the textbook Artificial Intelligence: A Modern Approach (1995), which has held a key place in university-level AI education for several decades, as a reference for this discussion. We also seek to highlight the perhaps less evident potential of an approach that values the conceptual and historical evolution of the field.