AI "Boom" or AI "Doom"? Analyzing Narratives in U.S. State-Level Policy

• In Event: AI Policy, Law, and Regulatory Diffusion

Abstract

In the absence of comprehensive federal legislation, most states in the U.S. have proposed and enacted local AI-related bills. Over the years, this process has created a complex emergent legal landscape consisting of over 1,000 AI state-level bills introduced since January 2023. This diverse topography reveals how political decision-makers interpret a new technology that is rapidly changing and widely misunderstood. At the same time, the language policymakers choose is never neutral, but rather signals which futures are deemed desirable, which populations are presumed vulnerable, and which risks are worth regulating. This paper analyzes these policies as sociological artifacts that encode dominant cultural narratives about technology, risk, and social order. Rather than evaluating these bills for their technical accuracy or policy effectiveness, we treat them as discursive objects worthy of analysis in their own right. Drawing on a corpus of state-level AI bills and their accompanying statements of purpose sourced from Legiscan, we employ natural language processing methods, including sentiment analysis and topic modeling, to examine the distributions of two policy narratives: "AI boom" (opportunity, economic competitiveness) and "AI doom" (risk, social disruption, precaution). This corpus spans legislation addressing algorithmic accountability, automated decision-making, generative AI, facial recognition, and data privacy, among other domains. We further investigate whether narrative patterns correlate with lobbying efforts, geographic region or state political affiliation, determining whether these factors are a strong predictor of narrative framing. Grounded in sociotechnical theory and the sociology of law, this study analyzes policy as a performative act that helps shape the collective understanding of AI. Policy can be used both as a line of defense against a novel technology viewed as a threat to social balance, or as a catalyst for its integration into society. By mapping the affective and ideological profiles of AI governance across U.S. states, we seek to capture what American policymakers actually fear, value, and imagine when they legislate the future of artificial intelligence.

Author

• Emily Li Hong, Brown University
  Presenting Author