Ethics of E-Learning Recommender Systems: Epistemic Positioning and Ideological Orientation

Abstract

AbstractRecommender systems are increasingly used in e-learning to provide users with personalized services and advice. Depending on the specific context for which the system is implemented (e.g., homework on a specific subject for university students, new training courses for life-long learners), the objectives and proposed items, the chosen recommendation techniques, the features that are considered, the way the recommendations are presented to the users are closely related to the designers’ perception of learners and knowledge. The various approaches reflect different epistemic and ethical viewpoints; for example, representing people using fixed models is easier to process, diagnose, predict and explain, but presents a partial view of reality and obscures the fact that they are complex and evolving individuals. Similarly, some filtering methods can restrict the view of available courses to items considered similar to those that the learner has already followed, thus promoting specialization rather than diversification and openness. This aspect is closely related to fundamental issues involved in the theory of knowledge, questioning the notions of utility and purposes of science, as well as a key issue for academic change and, more fundamentally, that of modern societies. Indeed, these issues should be seen in a broader context of reflection about the economic changes and ideological transformations of a society grounded on neoliberal capitalism. The main goal of this study is to explain how the design of recommender systems in e-learning has both ethical and practical implications since it reflects an ideological conception of science and techniques, thus requiring a previous examination of these issues in order to define the theoretical model of knowledge in which it takes place. For that purpose, we study the certain visions of teaching and learning that can be brought about by algorithms and models used by existing recommender systems in e-learning.