AI Development for the Public Interest: From Abstraction Traps to Sociotechnical Risks
This work addresses the lack of graduate-level interventions for integrating ethical and social contexts into AI education, aiming to unify pedagogical approaches across currently siloed AI subfields.
This paper explores the historical development of sociotechnical inquiry within AI Safety, Fair Machine Learning, and Human-in-the-Loop Autonomy. It reveals how each subfield's understanding of Public Interest Technology (PIT) is shaped by past challenges in integrating technical systems into social norms, and how these histories influence their responses to conceptual traps.
Despite interest in communicating ethical problems and social contexts within the undergraduate curriculum to advance Public Interest Technology (PIT) goals, interventions at the graduate level remain largely unexplored. This may be due to the conflicting ways through which distinct Artificial Intelligence (AI) research tracks conceive of their interface with social contexts. In this paper we track the historical emergence of sociotechnical inquiry in three distinct subfields of AI research: AI Safety, Fair Machine Learning (Fair ML) and Human-in-the-Loop (HIL) Autonomy. We show that for each subfield, perceptions of PIT stem from the particular dangers faced by past integration of technical systems within a normative social order. We further interrogate how these histories dictate the response of each subfield to conceptual traps, as defined in the Science and Technology Studies literature. Finally, through a comparative analysis of these currently siloed fields, we present a roadmap for a unified approach to sociotechnical graduate pedagogy in AI.