Search

Technological advances are enabling roles for machines that present novel ethical challenges. The study of 'AI ethics' has emerged to confront these challenges, and connects perspectives from philosophy, computer science, law, and economics. Less represented in these interdisciplinary efforts is the perspective of cognitive science. We propose a framework – computational ethics – that specifies how the ethical challenges of AI can be partially addressed by incorporating the study of human moral decision-making. The driver of this framework is a computational version of reflective equilibrium (RE), an approach that seeks coherence between considered judgments and governing principles. The framework has two goals: (i) to inform the engineering of ethical AI systems, and (ii) to characterize human moral judgment and decision-making in computational terms. Working jointly towards these two goals will create the opportunity to integrate diverse research questions, bring together multiple academic communities, uncover new interdisciplinary research topics, and shed light on centuries-old philosophical questions. publishedVersion

COVID-19 was a major shock to youth entrepreneurs and their businesses in Kenya. This paper studies the causal impact of grants—worth two months of baseline business revenue—and business development services as potential mitigation measures. Using multiple rounds of phone surveys up to seven months from the start of the pandemic, the analysis finds that youth who are assigned business grants or a combination of grants and business development services are significantly more likely to maintain a business, earn more revenue and profits, retain employees, and report higher confidence and satisfaction with life. There are no corresponding effects of business development services alone, although the follow-up period is extremely short for training effects to materialize. These results suggest that cash infusion for young entrepreneurs in times of an aggregate shock can be instrumental in moderating its immediate harmful impacts.

Long-acting reversible contraceptives are highly effective in preventing unintended pregnancies, but take-up remains low. This paper analyzes a randomized controlled trial of interventions addressing two barriers to long-acting reversible contraceptive adoption, credit, and informational constraints. The study offered discounts to the clients of a women’s hospital in Yaoundé, Cameroon, and cross-randomized a counseling strategy that encourages shared decision-making using a tablet-based app that ranks modern methods. Discounts increased uptake by 50 percent, with larger effects for adolescents. Shared decision-making tripled the share of clients adopting a long-acting reversible contraceptive at full price, from 11 to 35 percent, and discounts had no incremental impact in this group.

Covid-19 and the measures taken to contain it have led to unprecedented constraints on work and leisure activities, across the world. This paper uses nationally representative surveys to document how people of different ages and incomes have been affected across six countries (China, South Korea, Japan, Italy, UK and US). We first document changes in income/work and leisure. Second, we document self-reported negative and positive non-financial effects of the crisis. We then examine attitudes towards recommendations (wearing a mask in particular) and the approach taken by public authorities. We find similarities across countries in how people of different generations have been affected. Young people have experienced more drastic changes to their lives, and overall they are less supportive of these measures. These patterns are less clear across income groups: while some countries have managed to shield lower income individuals from negative consequences, others have not. We also show that how people have been affected by the crisis (positively or negatively) does little to explain whether or not they support measures implemented by the public authorities. Young people are overall less supportive of such measures independently of how they have been affected.

Disease spread is in part a function of individual behavior. We examine the factors predicting individual behavior during the Covid-19 pandemic in the United States using novel data collected by Belot et al. (2020). Among other factors, we show that people with lower income, less flexible work arrangements (e.g., an inability to tele-work) and lack of outside space at home are less likely to engage in behaviors, such as social distancing, that limit the spread of disease. We also find evidence that region, gender and beliefs predict behavior. Broadly, our findings align with typical relationships between health and socio-economic status. Moreover, they suggest that the burden of measures designed to stem the pandemic are unevenly distributed across socio-demographic groups in ways that affect behavior and thus potentially the spread of illness. Policies that assume otherwise are unlikely to be effective or sustainable.

This paper presents a new data set collected on representative samples across 6 countries: China, South Korea, Japan, Italy, the UK and the four largest states in the US. The information collected relates to work and living situations, income, behavior (such as social-distancing, hand-washing and wearing a face mask), beliefs about the Covid 19 pandemic and exposure to the virus, socio-demographic characteristics and pre-pandemic health characteristics. In each country, the samples are nationally representative along three dimensions: age, gender, and household income, and in the US, it is also representative for race. The data were collected in the third week of April 2020. The data set could be used for multiple purposes, including calibrating certain parameters used in economic and epidemiological models, or for documenting the impact of the crisis on individuals, both in financial and psychological terms, and for understanding the scope for policy intervention by documenting how people have adjusted their behavior as a result of the Covid-19 pandemic and their perceptions regarding the measures implemented in their countries. The data is publicly available.

Working with the Mexican Ministry of Education, this study piloted a scalable program to reduce high school dropout rates by focusing on socio-emotional skill development and mathematics tutoring. The intervention was evaluated through a randomized field experiment with more than 5,000 youths at 20 upper secondary schools in Mexico City. An intention-to-treat analysis finds some evidence that exposure to the Opportunities and Development to Avoid Risks Program increases socio-emotional skills, but no evidence that it improves math outcomes or future attendance. Likely explanations for these 0 results include low take-up and other process factors, which are document qualitatively, as well as heterogeneous treatment effects. In particular, an inverse-probability-weighted matching model is suggestive of an effect whereby some students participate actively in the program and drop out of school less often, while other students choose not to participate when given the option and actually drop out more as a result.

Bureaucratic performance is a crucial determinant of economic growth, but little real-world evidence exists on how to improve it, especially in resource-constrained settings. We conducted a field experiment of a social recognition intervention to improve record keeping in health facilities in two Nigerian states, replicating the intervention - implemented by a single organization - on bureaucrats performing identical tasks. Social recognition improved performance in one state but had no effect in the other, highlighting both the potential benefits and also the sometimes-limited generalizability of behavioral interventions. Furthermore, differences in facility-level observables did not explain cross-state differences in impacts, suggesting that it may often be difficult to predict external validity.

Bureaucratic performance is a crucial determinant of economic growth. Little is known about how to improve it in resource-constrained settings. This study describes a field trial of a social recognition intervention to improve record keeping in clinics in two Nigerian states, replicating the intervention -- implemented by a single organization -- on bureaucrats performing identical tasks in both states. Social recognition improved performance in one state but had no effect in the other, highlighting both the potential and the limitations of behavioral interventions. Differences in observables did not explain cross-state differences in impacts, however, illustrating the limitations of observable-based approaches to external validity.

Although the concept of randomized assignment to control for extraneous factors reaches back hundreds of years, the first empirical use appears to have been in an 1835 trial of homeopathic medicine. Throughout the 19th century, there was primarily a growing awareness of the need for careful comparison groups, albeit often without the realization that randomization could be a particularly clean method to achieve that goal. In the second and more crucial phase of this history, four separate but related disciplines introduced randomized control trials within a few years of one another in the 1920s: agricultural science, clinical medicine, educational psychology, and social policy (specifically political science). Randomized control trials brought more rigor to fields that were in the process of expanding their purviews and focusing more on causal relationships. In the third phase, the 1950s through the 1970s saw a surge of interest in more applied randomized experiments in economics and elsewhere, in the lab and especially in the field.

Laboratory experiments were conducted to investigate how private and public information affect the selection of an innovation and the timing of adoption. The results shed light on the behavioral anomaly called the “energy-efficiency gap” in which consumers and firms delay adoption of cost-effective energy and environmental innovations. The subjects chose between competing innovations with freedom to select the timing of their adoption, relying on private signals and possibly observation of their peers. When deciding whether to make an irreversible choice between safe and risky technologies, roughly half the subjects delayed adoption beyond the time indicated by equilibrium behavior -- confirming the behavioral anomaly found for environmental innovations. When they did adopt, the subjects gave proportionately more weight to their private signals than to the actions of their peers, implying they do not ‘herd’ on the latter. Nevertheless, when the subjects observed their peers’ decisions, they did accelerate the timing of their adoption despite not necessarily imitating their peers. This result occurred even when the payoffs were statistically independent, as if observing prior adoptions exerted ‘peer pressure’ on the subjects to act. The experimental results suggest that rapid dissemination of information about peer actions can speed up the diffusion of environmental innovations and improve selection among competing technologies.

The paper shows that self-control, time preferences, and values are malleable in adults, and that investments in these skills and preferences reduce crime and violence. The authors recruited criminally-engaged Liberian men and randomized half to eight weeks of group cognitive behavioral therapy, fostering self-regulation, patience, and noncriminal values. They also randomized $200 grants. Cash alone and therapy alone dramatically reduced crime and violence, but effects dissipated within a year. When cash followed therapy, however, crime and violence decreased by as much as 50 percent for at least a year. They hypothesize that cash reinforced therapy's lessons by prolonging practice and self-investment.