AI and Developing Socially-Engaged Computational Thinkers

This position paper reflects on why we want greater participation in the computing sciences, what ‘greater participation’ means and should mean, and on the particular characteristics of AI that recommend it as an enticement and way into computing. Primary tenets are that the need for and training of socially-engaged ‘computational thinkers’ (Wing, 2006) is the most important motivation for increasing the diversity (and numbers) of students in the computing sciences and that ‘AI for the social good’ may bring students into computing for whom AI per se would not. Computational Thinkers In this time of rapid technological and societal change, when policy and practice involving technology have enormous consequences (e.g., global climate change), often looming beyond some myopically-construed horizon, we need what Jeannette Wing (2006) calls “computational thinkers”, which I take as critical thinkers who are conversant with computational principles, tools, practices, and broader impacts. Wing’s desiderata for human thinkers certainly includes deterministic skills of comparison, tracing, and simulation, but her vision is dominated by a desire that human problem solvers exhibit a sophisticated heuristic and systematized approach to non-deterministic reasoning, aspects of thinking that AI seeks to codify. Socially-engaged computational thinkers are motivated to and capable of modeling and tracing through the societal and environmental consequences of technical and social designs and actions. Dismantling Compartmentalization The goal of training socially-engaged, computational thinkers is paramount. Professional societies and accrediting boards, most notably IEEE and ABET, are insisting on “an ability to analyze the local and global impact of computing on individuals, organizations, and Copyright © 2008, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved society.” (ABET criteria, p. 19) and like characteristics among computing graduates. These institutions recognize that segregating technical training from technology’s broader impacts do not serve society well. Diversity, not numbers per se It is fortuitous that dismantling the compartmentalization between computing and larger societal purposes is synergistic with broadening the diversity of recruits to computing. Studies show that many women, for example, are drawn into computing when they see its fit into larger societal purposes [Margolis & Fisher, 2002]. In this case, greater diversity reinforces a greater community vision for computing’s utility, and vice versa. The diversity goes beyond gender, race, and culture. Under the assumption that computational competence is a necessity, diversity of values (or of value weighting) is vital.