Research

One of the promising opportunities offered by AI to support the decarbonization of electricity grids is to align demand with low-carbon supply. We evaluated the effects of one of the world’s largest AI managed EV charging tariffs (a retail electricity pricing plan) using a large-scale natural field experiment. The tariff dynamically controlled vehicle charging to follow real-time wholesale electricity prices and coordinate and optimize charging for the grid and the consumer through AI. We randomized financial incentives to encourage enrollment onto the tariff. Over more than a year, we found that the tariff led to a 42% reduction in household electricity demand during peak hours, with 100% of this demand shifted to low-cost, low-emission off-peak periods. The tariff generated substantial consumer savings, while demonstrating potential to lower producer costs, energy system costs, and carbon emissions through significant load shifting. Overrides of the AI algorithm were low, suggesting that this tariff is likely more efficient than a real-time-pricing tariff without AI. Our findings highlight the potential for scalable AI managed charging and its substantial welfare gains for the electricity system and society. We also show that experimental estimates differed meaningfully from those obtained via non-randomized difference-in-differences analysis, due to differences in the samples in the two evaluation strategies, although we can reconcile the estimates with observables.

Exposure to fine particulate matter (PM2.5) poses major health risks, especially in rapidly urbanizing cities. As urbanization accelerates, people in low- and middle-income countries spend more time indoors, where pollution risks remain poorly understood. We present evidence from over 152,000 monitor-hours of indoor PM2.5 measurements across homes in Jakarta, Indonesia, one of the world’s largest and most polluted cities. We find that mean daily indoor and outdoor PM2.5 levels are both dangerously high, eight times above World Health Organization’s (WHO) health-based guidelines. In addition, indoor PM2.5 frequently reach hazardous levels—40 to 100 times the WHO guideline, levels that outdoor monitors do not capture. Unlike in developed settings, most indoor pollution originates from outdoor infiltration. Survey data also reveal large inequalities: lower-income households experience double the mean indoor PM2.5 of higher-income households. Our findings show that indoor air pollution remains both severe and unequally distributed, even in this population where most people have adopted cleaner cooking fuels. Researchers and policymakers should integrate outdoor air quality mapping with demographically representative indoor monitoring to close key data gaps, enabling more accurate exposure estimates and better-targeted environmental health policies.

Indonesian wildfires generated an average of 1.5 million tons of PM2.5 per year from 2003 to 2018—1.5 times the emissions of U.S. wildfires. Fires’ connection to agriculture complicates impact assessments, as pollution and income from agricultural activities jointly shape health outcomes. By leveraging random variations in wind direction and fire locations, we disentangle the pollution and income effects. Upwind fires—those carried by the wind toward a region—serve as an instrument to isolate the pollution impact of agricultural fires. A 10 µg/m³ increase in PM2.5 reduces birthweight by 20.5 grams—equivalent to the impact of smoking five cigarettes daily during pregnancy—and lowers the probability of a child being developmentally on track at ages 3–5 by 5%. In contrast, local fires show no significant effect on child outcomes, emphasizing the confounding influence of economic activity. Furthermore, Indonesia’s conditional cash transfer (CCT) program offsets the adverse health impacts of pollution, demonstrating how financial support can enable protective responses. (Draft available upon request)

Climate change and air pollution are closely intertwined and share root causes — mainly fossil fuel combustion. Air pollution, with its immediate and visible health impacts can serve as a powerful entry point for broader environmental action. Yet much of the world lacks local pollution monitoring, limiting public awareness and pressure for reform. This project will run the first global randomized control trial to test whether delivering real-time, localized air quality data to stakeholders can raise awareness, prompt engagement, and reduce pollution. The project leverages novel, satellite-derived PM2.5 estimates to track daily air quality across 1,000+ cities and send alerts via social media to government officials, regulators, NGOs, journalists, and citizens in treated cities. This low-cost, scalable intervention offers the potential to bridge environmental data gaps and strengthen accountability around the world. Providing transparent information on pollution exposure can empower communities to demand and shape effective responses to air pollution and climate change.