Global Environmental Policy Seminar with Matto Mildenberger

When Machines Read Politics: Agentic AI, Political Experts,and Measuring Energy Transition Policymaking Windows across 25 Countries What explains cross-national variation in the political economy of the energy transition? Systematic comparisons of transition politics requires at scale measurements of hard-to-observe political factors, including policy support by legislators, political parties, national elites, courts, firms, financial regulators and the media, and how those commitments shift over time. In this talk, I will share results from a research programme that uses scaled agentic AI systems to collect, process and classify millions of parliamentary speeches, party manifestos, court decisions, media reports and regulatory records across 25 countries. The project also leverages in-depth qualitative human expert assessments of national-level political economy forces, capturing power relations absent from textual records. A Bayesian measurement model combines these diverse qualitative and quantitative empirical datastreams into per-country estimates of when and where energy policymaking windows are opening or closing. I will also reflect on when AI assessments of political power appear most accurate, and how AI and expert assessments of political economy can be combined to scale our understanding of the clean energy transition. Bio: Matto Mildenberger is Associate Professor of Political Science at the University of California Santa Barbara, where he also directs UCSB's 2035 Initiative, a campus-wide think-and-do tank developing applied, climate and energy policy roadmaps. Matto's research focuses on the political obstacles to climate and energy policy action and how both elites and publics make sense of climate policy.