Scaffolding Human Champions: AI as a More Competent Other

Abstract

AbstractArtificial intelligence (AI) has surpassed humans in a number of specialised intellectual activities—chess and Go being two of many examples. Amongst the many potential consequences of such a development, I focus on how we can utilise cutting edge AI to promote human learning. The purpose of this article is to explore how a specialised AI can be utilised in a manner that promotes human growth by acting as a tutor to our champions. A framework for using AI as a tutor of human champions based on Vygotsky’s theory of human learning is here presented. It is based on a philosophical analysis of AI capabilities, key aspects of Vygotsky’s theory of human learning, and existing research on intelligent tutoring systems. The main method employed is the theoretical development of a generalised framework for AI powered expert learning systems, using chess and Go as examples. In addition to this, data from public interviews with top professionals in the games of chess and Go are used to examine the feasibility and realism of using AI in such a manner. Basing the analysis on Vygotsky’s socio-cultural theory of development, I explain how AI operates in the zone of proximal development of our champions and how even non-educational AI systems can perform certain scaffolding functions. I then argue that AI combined with basic modules from intelligent tutoring systems could perform even more scaffolding functions, but that the most interesting constellation right now is scaffolding by a group consisting of AI in combination with human peers and instructors.