Policy

Predictable Disasters: AI & The Future of Crisis Response, Brookings, December 2021 (with Grant Gordon)

The greatest barrier to achieving many of the Sustainable Development Goals (SDGs) lies in fragile settings characterized by extreme poverty, weak institutions, and ongoing vulnerability to natural and man-made disasters. Given current trends, complex emergencies may become even more challenging over the next decade, however, artificial intelligence (AI) holds the potential to transform crisis response to both save and improve many lives.  In order to realize that promise, crisis response policymakers will have to prioritize ongoing and new AI investments based on a sophisticated understanding of risk and return. 

Bending the curve to meet the SDGs in fragile settings will require new tools and radical improvements in the impact, scalability, or cost-effectiveness of current practices – an ambitious goal that can be supported by the exponential growth in promising machine learning applications. In turn, harnessing AI for crisis response requires a clear-eyed understanding of the conditions under which machine learning can improve outcomes as well as a framework for when and how to effectively integrate it into organizations. 

Full text online: https://www.brookings.edu/wp-content/uploads/2021/12/Chapter-Six_Breakthrough.pdf

The Sting in COVID-19's Tail, Foreign Affairs, January 2021

At the outset of the COVID-19 pandemic, many feared that the disease would hit the world’s poorest countries the hardest—that cases would overwhelm hospitals, health-care workers would run out of equipment and supplies, and the death toll would be devastating. Those eventualities, thankfully, have not yet come to pass. From available data, there appear to be fewer deaths per capita in low- and middle-income countries than in their richer counterparts.

But developing countries may have suffered more from the pandemic economically and politically than they have in the realm of public health. A recent International Monetary Fund (IMF) study has found that social unrest increased after the SARS, H1N1, MERS, Ebola, and Zika pandemics, particularly in countries with high levels of inequality. COVID-19 has more potential than all those past diseases combined to increase poverty, deepen social fractures, and intensify conflicts. 

Full text online: https://www.foreignaffairs.com/articles/world/2021-01-26/sting-covid-19s-tail 

Climate Change Doesn't Have to Stoke Conflict, Foreign Affairs, September 2020 (with Rob Malley) 

Among the many sobering projections of harm to be caused by climate change is this eye-popping statistic: on average, according to economists, a rise in local temperature of half a degree Celsius is associated with a ten to 20 percent increase in the risk of deadly conflict. If accurate, that means the likelihood of such strife is swiftly rising. UN climate scientists estimate that manmade emissions have generated one degree of global warming since preindustrial times, and because the pace of climate change is fast accelerating, they predict another half a degree of warming as soon as 2030. Tropical areas will have even more extreme warming, with a correspondingly higher risk of climate-related insecurity.  

Ending or preventing conflicts exacerbated by climate change requires a faster and different approach than addressing climate change itself. Many governments have begun to curb emissions, but they are gradually phasing in their climate mitigation efforts. For example, 120 countries have embraced a net-zero carbon emission target by 2050—a worthy goal that could prevent the earth from eventually becoming uninhabitable. But millions of people around the world are already experiencing record heat waves, extreme precipitation, and rising sea levels—changes that disrupt livelihoods; exacerbate food insecurity, water scarcity, and resource competition; and spur migration. Tackling climate change is a necessary but inescapably longer-term endeavor. Conflict prevention must happen now. 

Full text online: https://www.foreignaffairs.com/articles/ethiopia/2020-09-28/climate-change-doesnt-have-stoke-conflict 

How Employers in Poor Countries Are Using Nudges to Help Employees Save Money, Harvard Business Review, June 2018 (with Joshua Blumenstock and Michael Callen)

One of the most common ways to get people to save is through their employer. In particular, behavioral economics—that marriage of economics and psychology that has put terms like “nudge” into the popular lexicon—has provided a powerful tool for increasing savings, in the form of the default enrollment. The idea is simple: people save more in retirement accounts when they are automatically enrolled by their employer than when they have to sign up themselves. Most U.S. employers have adopted default savings programs, but the idea is only just entering  poor countries, where saving is less common. According to World Bank figures, half of adults in high-income OECD countries save in a formal account; in developing economies, it’s only one in five. But we now have some evidence about how to scale nudges to help change this.

Along with other aspects of the formal financial ecosystem—ubiquitous banks, ATMs, and credit cards—default savings programs have historically been the province of rich countries. But recently, the world has seen a dramatic drop in the numbers of unbanked adults due to the rapid spread of mobile money. The number of people with phones is far higher than the number of people who have easy access to traditional bank branches, especially among the rural poor. Prior to our study, however, no one had tested whether mobile banking could facilitate default savings programs. In a recent paper, we found that nudging employees to save worked in Afghanistan, one of the world’s least financially developed countries.

Full text online: https://hbr.org/2018/06/how-employers-in-poor-countries-are-using-nudges-to-help-employees-save-money