Beatrice Magistro

📖New book forthcoming in June 2026 from
@caltechlcssp.bsky.social Research Affiliate, @beamagistro.bsky.social!

Who Thinks Like an Economist? How the Economist Mental Model Shapes Political Decisions

Pre-order it here from @cambridge.org: cambridge.org/us/universit...

Conceptual photo illustration of two people looking at a robot. The man looks intrigued, while the woman looks disgusted. Stock photo.

Women think AI is riskier than men do, according to a survey. Women are more risk averse in general, and may be more occupationally threatened by workplace automation. In PNAS Nexus: https://ow.ly/OakG50YcqsR

Nice! This aligns well with recent reporting by @sarahoconnorft.ft.com in the FT The AI Shift: why do men use AI at work more than women? - giftarticle.ft.com/giftarticle/...

6/ The bottom line: whether or not AI delivers on its promises, differential adoption and policy preferences may create new axes of inequality.

5/ This matters for governance too. Women are more supportive of government intervention to slow AI adoption, especially when job outcomes are risky.

4/ We ran a survey experiment varying the probability that AI adoption would create more jobs than it destroys. Key finding: when job gains are uncertain (30% probability), women's support is ~13% lower than men's. When gains are certain (100%), the gap disappears.

3/ Two mechanisms: (1) women have higher general risk aversion, and (2) women face greater exposure to AI-related employment risks due to occupational segregation.

2/ Women are more skeptical of AI than men. In earlier work, we documented this gap but couldn't fully explain it. This paper digs into why and we explore the role of risk.

Explaining women’s skepticism toward artificial intelligence: The role of risk orientation and risk exposure Abstract. This article examines the gender gap in attitudes toward the adoption of AI in the workplace, with a focus on how gender differences in risk orie

1/ 🚨 New paper out today in @pnasnexus.org w/ @sborwein.bsky.social @rmichaelalvarez.bsky.social @bartbonikowski.bsky.social and Peter Loewen: "Explaining Women's Skepticism toward Artificial Intelligence: The Role of Risk Orientation and Risk Exposure" 🧵
academic.oup.com/pnasnexus/ar...

Our takeaway: policymakers hoping to avoid a protectionist backlash against AI should act now.
Citizens' beliefs about the economy (not just their material conditions) drive political behavior. The window for pre-emptive policy may be closing.

What makes AI different from past automation shocks?
- It threatens cognitive workers (not just routine/manual)
- The scale of disruption raises the stakes
- Politicians learned from trade that mobilizing backlash works

These alignments exist before any party has claimed AI as an issue. The raw material for politicization is already there, waiting for political entrepreneurs to exploit it.

Policy preferences differ too.
Substituters favor policies that delay job loss: taxing AI, regulating AI, reducing immigration.
Complementers favor policies that help workers adapt: investing in education, reskilling programs.

Causal beliefs predict vote choice.
🇺🇸 In the US: Substituters → more likely Republican; Complementers → more likely Democrat
🇨🇦 In Canada: Substituters → more likely NDP; Complementers → more likely Liberal

We found 4 types:
🔴 Substituters (37.5%): AI replaces workers, harms consumers, only benefits firms
🟢 Complementers (33.5%): AI augments skills, mostly benefits firms and consumers
🟡 Uncertains (17%): Don't know
⚪ Skeptics (12%): AI won't change much

We asked people to reason through AI's effects on:
- Company profits & demand
- Product quality & prices
- Workers' skills, wages & hiring
Then we used latent class analysis to identify distinct "causal theories" of AI in the public.

We surveyed 6,000 🇺🇸 and 🇨🇦 to understand how people think about AI's economic effects, and whether these beliefs could fuel political backlash

1/ 🚨 New paper ("Causal Beliefs and the Potential for Political Backlash Against AI") accepted at @poqjournal.bsky.social w/ @sborwein.bsky.social @rmichaelalvarez.bsky.social @bartbonikowski.bsky.social and Peter Loewen.
🧵👇
doi.org/10.31219/osf...

Can retraining programs ease the fear of AI job loss? Public policy expert provides new insights New findings from a study show that Americans and Canadians want the government to address AI-related job loss through workplace retraining.

Americans and Canadians overwhelmingly agree that employees need more training in order to harness the power of AI and prevent it from stealing their jobs, regardless of political leaning, according to a multiyear study performed by Prof @beamagistro.bsky.social

Appealing to Independents: information on negative externalities increases support for environmental corrective taxes Climate change requires urgent global action, but efforts to implement solutions like a carbon tax face deep political polarization, particularly in the United States. This study explores how fram...

Appealing to Independents: information on negative externalities increases support for environmental corrective taxes by @beamagistro.bsky.social & @rmichaelalvarez.bsky.social.

doi.org/10.1080/0964...

The Coming AI Backlash How the anger economy will supercharge populism.

New in @foreignaffairs.com w/ @sborwein.bsky.social @rmichaelalvarez.bsky.social @bartbonikowski.bsky.social and Peter Loewen.

AI is already transforming labor markets, and governments risk repeating the mistakes they made with globalization

www.foreignaffairs.com/united-state...

It’s been two years since I’ve seen @beamagistro.bsky.social in person but tomorrow we are changing that and I couldn’t be more excited. This is news the internet needs.

A Conversation with Beatrice Magistro: AI, Climate Change, & her book, Who Thinks Like an Economist YouTube video by Linde Center for Science, Society, and Policy

Listen to our latest LCSSP video as Co-Director @rmichaelalvarez.bsky.social visits with @caltech.edu post-doc scholar @beamagistro.bsky.social, Research Associate in Computational Social Science, about their joint work, her research in AI and Climate change & her new book!
youtu.be/6kkSNTesvXM?...

Political ideology and views toward solar geoengineering in the United States Political polarization remains a significant barrier to effective climate action in the United States. Conservatives often express skepticism toward climate change policies emphasizing government inte...

New research by @beamagistro.bsky.social and colleagues suggests that increasing public awareness of solar geoengineering could foster bipartisan engagement with climate policy in the US

journals.plos.org/climate/arti...

@rd545.bsky.social @paulowennberg.bsky.social @rmichaelalvarez.bsky.social

Book Project Who Thinks Like an Economist? How the Economist Mental Model Shapes Political DecisionsForthcoming at Cambridge University Press Introduction Available HereRecent populist waves—from Donald Trump’s…

13/ Full introduction and ToC available here 👉 beatricemagistro.com/research/boo... stay tuned for the book and happy to share any chapters 🙂

12/ You can't beat zero-sum populism with fact-checks alone. Where the EMM is scarce, zero-sum narratives and economic populist appeals find more fertile ground.

11/ Individuals who think like economists are 1) more likely to favor welfare-enhancing policies (like trade and immigration), 2) far less zero-sum, and 3) responsive to economic information: they update when cost-benefit data change, while AMM voters stick to partisan cues.

10/ I build new indices of economic knowledge that identify who's using the EMM (it's more than just education). Then I test them with surveys, conjoint experiments, and a classroom study across Italy, the UK, and the US, spanning trade, immigration, price controls, and AI. 🌍

9/ This divergence in mental models—how we think, not what we value—drives wildly different readings of political events and policy choices.