How Citizens Perceive Local AI Data Center Development: Economic Benefits, Environmental Tradeoffs, and Policy Framing

• In Event: The Infrastructure of Intelligence: Navigating Siting Conflicts, Community Equity, and Grid Resilience in the AI Era

Abstract

As artificial intelligence (AI) expands rapidly across sectors of the economy, AI data centers are emerging as a consequential form of local infrastructure development. These facilities promise economic benefits, including tax revenue, employment opportunities, and community investment, but they also generate concerns related to electricity demand, water consumption for cooling, and siting near residential areas. Despite increasing policy attention to digital infrastructure expansion, there is limited empirical evidence on how citizens evaluate tradeoffs associated with proposed AI data center projects or whether support depends on how such projects are framed. This study addresses that gap through a framing-embedded conjoint survey experiment examining public preferences for local AI data center development. The study draws on an original survey of approximately 1,100 adults in a representative sample of residents across U.S. Midwestern states. The Midwest is increasingly identified as a strategic destination for AI data center expansion due to relatively abundant wind and solar resources, comparatively low water scarcity, and favorable projected carbon and water intensities. As these states absorb a growing share of proposed AI infrastructure investment, understanding how residents evaluate associated economic and environmental tradeoffs becomes an important policy question for state and local governments. To examine these preferences, respondents first received a brief definition of an AI data center and were randomly assigned to one of three framing conditions: technological progress, regional competition, or a neutral baseline describing a local development proposal. Respondents then completed a conjoint experiment in which they evaluated four paired comparisons of hypothetical mid-sized AI data center projects varying across six policy-relevant attributes: projected local tax revenue, permanent jobs, annual community benefit funding, water use for cooling, expected household utility bill impacts, and distance from residential areas. For each pair, respondents selected their preferred project and rated each proposal on a 1–7 approval scale, enabling analysis of both comparative preferences and absolute levels of support. Preliminary analyses suggest that respondents place the greatest weight on direct household-level impacts, particularly anticipated utility bill changes and proximity to residential areas, while also valuing fiscal benefits such as tax revenue and employment creation. Framing projects in terms of technological progress or interregional competition is expected to modestly influence support overall, and framing effects may vary across partisan groups and between rural and urban residents. The study makes two contributions. First, it identifies which attributes of AI data center proposals most strongly shape public support, clarifying how residents balance economic development benefits against infrastructure-related burdens. Second, it evaluates whether broader interpretive frames alter these preferences by testing two narratives increasingly prominent in policy debates over AI infrastructure: technological modernization and interlocal economic competition. The findings contribute to emerging research on AI governance and infrastructure siting politics and provide policymakers with evidence on how communities evaluate proposed AI data center projects as digital infrastructure becomes an increasingly visible component of local development strategy.

Author

• Guimin Zheng, Iowa State University
  Presenting Author