A classification framework for generative artificial intelligence for social good

Abstract

Many policy makers and corporate leaders are adjusting their strategies to harness the power of GenAI. There are numerous debates on how GenAI would fundamentally change existing business models. However, there is not much discussion on roles of generative AI in the domain of social good. Broader views covering potential opportunities of GenAI to enable diverse initiatives in the social good space are largely missing. We intend to reduce the gap by developing a classification framework that should allow researchers gauge the potential impact of GenAI for social good initiatives. Through case analysis, we assess how value-added abilities of GenAI may influence various social good initiatives. We adopt/develop two loosely connected classification frameworks that are grounded in task-technology fit (TTF) theory. Subsequently, we investigate how our analyses of GenAI initiatives utilizing different dimensions of these two frameworks may be synthesized to provide appropriate explanation for potential success of GenAI for social good. We develop five propositions that will provide guidance to practitioners and researchers. The theoretically grounded analysis of 21 GenAI for social good use cases based on the two classification frameworks, and the resulting propositions are the original contributions of this paper to the AI for social good literature.