Research

How will Artificial Intelligence impact politics? My dissertation investigates how economic shocks, especially those driven by automation and AI, alter public preferences for redistribution, regulation, and the welfare state. Some economic shocks provoke populist backlash; others are met with political stability.

I use experimental and cross‑national observational methods to examine when and why these divergent responses occur. Latin America, where welfare regimes vary widely and populist politics are deeply rooted, is a key region of comparison in my work and helps explain why similar shocks can produce support for either left‑ or right‑wing populists. By situating AI within a broader comparative agenda on the politics of economic disruption, I aim to help policymakers manage AI’s political impacts.

Dissertation Projects

‘Compensate the AI Losers?’: A Video Experiment on AI Framing and Policy Preferences (Working Paper)

*Previously Titled “AI Winners vs. AI Losers: Experimental Evidence on Economic Risk & Policy Preferences”

What impact does messaging about AI have on policy preferences? Using a preregistered video survey experiment (N=1,002) of U.S. white-collar workers, I examine how messaging about AI as either a job threat or productivity tool shapes policy preferences on redistribution and AI regulations (predistribution). In the absence of AI-messaging, I find voters have baseline pessimism about AI’s wider impact, but are slightly optimistic about their own income. However, voters are far more responsive to AI-threat messaging than optimism frames. When faced with AI threat frames, those with less education prefer “predistribution” in the form of AI regulation over wealth redistribution. My findings suggest these preferences are motivated by increased sociotropic concerns about AI. Partisanship emerges as a key lens through which voters interpret the AI threat. Republicans respond to increased AI risk by rejecting social safety nets, but no significant results emerge for Democrats.

The Welfare State, Immigration, and Explaining the Success of Left Populists

Electorates are sometimes in “populist moods,” but vary in whether they elect left‑wing or right‑wing populists. This chapter proposes that welfare state generosity moderates this variation. Unlike traditional left‑right divisions, both left and right populists defend the welfare state but in different ways.

Left populists (e.g., Evo Morales) tend to advocate for protecting the welfare state from austerity.
Right populists (e.g., Jair Bolsonaro, Geert Wilders) often adopt “welfare chauvinism,” arguing that immigrants put too much pressure on the welfare state.

Using data from the Centro de Estudios Distributivos, Laborales y Sociales (CEDLAS), the Comparative Welfare Entitlements Dataset (CWED), and the IMF Banking Crises Database, I empirically test Dani Rodrik’s theoretical assertion that the nature of a globalization shock interacts with welfare state generosity to shape populist outcomes.

For example, an immigration shock in a country with a generous welfare state should produce support for right‑populist welfare chauvinists, while a financial shock in a less generous state should produce support for left‑populists arguing to protect the welfare state from austerity.

Can Governments Buy Support for Automation?

Governments are often called upon to regulate automation, from California’s 2024 self‑driving truck ban to its 2025 AI regulations. This project asks whether welfare state generosity can reduce public support for automation‑related regulations by drawing on the embedded liberalism literature, which argues that governments can “buy” support for disruptive economic changes by expanding welfare spending.

This project tests whether embedded liberalism applies to white‑collar workers facing AI‑based automation and whether they will accept workplace transformation with little political consequence, or instead punish politicians who accept automation. Using cross‑national regression analysis with CEDLAS, the General Social Survey, and CWED, this chapter shows how welfare state scope predicts whether voters want to restrict automation and helps policymakers avoid repeating the embedded liberalism breakdown of the 2010s.