Paper short abstract:

This paper explores how perceptions of AI’s nature and value, along with specific democratic ideals, shape the normative visions of citizens in current discussions on public participation in AI from perspectives in Public Understanding of Science, Public Engagement in Science, and Citizen Science.

Paper long abstract:

The recent proliferation of artificial intelligence (AI) tools has ignited numerous discussions about the role of citizens and the public amid the rapid development and implementation of AI. These discussions, centered around citizens and the public, are often seen as expressions of democratic values and practices. However, despite the trope of democratic engagement with AI, the portrayal of citizens in these conversations is often vague, implicit, and multifaceted. While frequently aimed at and dominated by AI designers, tech companies, and regulators, they have simultaneously projected normative imaginaries of ideal citizens. What constitutes “good citizens” within these conversations? And how do these characterizations relate to assumptions about the nature of AI technologies and the role of democracy in their development?

This paper offers a set of heuristics for interpreting the current discussions and practices surrounding public participation in AI. We draw on insights from over three decades of work in Public Understanding of Science (PUS), Public Engagement in/with Science (PES), and Citizen Science (CS) to analyze current debates about AI and public participation. It builds on how model citizens are envisioned in the scholarly traditions of PUS, PES, and CS, exploring how the imagined archetypes of literate (PUS), responsible (PES), and contributive citizens (CS) permeate contemporary debates about the public’s role in AI. Instead of presenting a systematic, comprehensive review, we will highlight typical examples that illustrate how perceptions of AI’s nature and value, combined with specific democratic ideals, shape corresponding normative visions of citizens.