On the one hand, more and more decisions are being made based on what machines can learn about us: who gets a loan, who gets into college, who gets insurance, etc. On the other hand, people have many reservations about the fairness of algorithms, about algorithmic perpetuation of biases built into historical data, about the mis- or overinterpretation of statistical correlations, and more. This grant funds work by economists Jens Ludwig from the University of Chicago and Sendhil Mullainathan from Harvard to study when, why, and how people should override recommendations based on artificial intelligence. The team will focus on how New York City judges decide to release or hold suspects before trial. Machine generated recommendations—ones that use facts about a suspect to predict whether that subject will commit a crime if released back into the community—are already in use. But judges are also privy to information about a subject that a typical algorithm is not, including a suspect’s courtroom dress, demeanor, accompanying associates, etc. Ludwig and Mullainathan will study whether and how these additional factors affect both judicial predictions of suspect behavior as well as AI predictions of judicial behavior.

This grant funds efforts by Avi Goldfarb, Joshua Gans, and Ajay Agrawal, three leading economists from the University of Toronto, and Catherine Tucker, Sloan Distinguished Professor of Management at MIT, to facilitate rigorous research on the economics of artificial intelligence (AI). Building on a successful conference on the economics of AI held in Toronto in 2017, the team plans to hold a series of three more annual conferences on related topics, commissioning papers for each conference, then publishing and disseminating the collected conference proceedings. Over three years, the team anticipates commissioning more than 50 academic papers. The team will also organize extensive training, support, and other services for graduate students and postdoctoral fellows interested in studying the economics of AI. The plan is to train more than 90 early-career researchers in advanced methodological and analytic techniques.

This grant provides partial support to the National Academy of Sciences for a consensus report on how to use insights from the behavioral, social, and statistical sciences to improve the reliability and reproducibility of research. The project will include five committee meetings featuring various experts across the sciences, five commissioned papers on reproducibility, an expert panel on behavioral economics and the professional incentives facing producers and consumers of research, and a final consensus report. While the majority of the funding will be provided by NSF, this grant will provide supplementary support both for including economic perspectives in the study and for disseminating the final report.  

Between 1947 and 1973, U.S. productivity grew an average of nearly 3 percent per year. Since 2007, that rate has dropped to 1.3 percent. Since 2010, it has plummeted to 0.5 percent. China and India excepted, other countries around the world have experienced similar drops in productivity growth. Why? What is going on? This grant funds a project by a team led by Louise Sheiner at the Hutchins Center on Fiscal and Monetary Policy at the Brookings Institution to shed light on this “productivity puzzle.” Over the next two years, Brookings will conduct and commission original and rigorous research on productivity, engage with stakeholders at U.S. statistical agencies about the quality and limitations of existing productivity measurements, produce 10 to 12 peer reviewed papers, hold six conferences on this and related issues, produce a conference volume, and disseminate recommendations on how to improve research and statistics about economic productivity.

This grant supports the launch and operation of a new working group at the National Bureau of Economic Research (NBER) dedicated to studying the “economics of science.” Led by Paula Stephan, the group will bring together top flight economists to share existing work and findings, identify new areas for research, examine methodological and data issues, and commission new research. Topics include incentives in the current system, how the structure of grants and review systems affects scientific risk taking, the costs and efficiencies of different research funding models, how to judge scientific quality, and how to measure return on investment in basic and applied science. Along with four meetings of the working group, the grant will fund administrative and planning costs, support for small research grants, and partnerships between the working group and institutions like research universities or other science funders.

This grant supports the expansion of a popular seminar on computational social science, run by Matthew Salganik of Princeton University and Christopher Bail of Duke University. The instructional program, which takes place over the summer, involves lectures, group problem sets, and participant-led research projects. The seminar also includes outside speakers who conduct computational social science research in academia, industry, and government. Topics covered include text as data, website scraping, digital field experiments, nonprobability sampling, mass collaboration, and ethics. Interest in the program has been robust, with more than 10 times as many applicants as available slots each year. Sloan funds will allow lectures and course content to be broadcast via interactive video to six new satellite locations, including City University of New York; Northwestern; University of Colorado, Boulder; Seattle; Helsinki; and Cape Town. Additional satellite sites may be added in future years.

Harvard econometrician Jose Zubizarreta is developing new statistical methods for the extraction of causal inferences from large datasets. His methods flexibly adjust for covariates in observational studies while also yielding more stable causal estimates. For part of the research, Zubizarreta will investigate formal and theoretical properties of these methods. His team, however, based as it is at a medical school, will also work on specific applications. These require, for example, developing a new framework for the design and analysis of observational studies with discontinuities, or developing new methods that improve the degree of control (covariate balance) and statistical efficiency of randomized experiments that enhance their generalizability. Zubizarreta plans to produce five peer-reviewed papers on these topics. In addition, all software, code, and examples will be produced in an open source programming language and made freely available, together with documentation and sample data, to the academic community and the public.

This grant funds an initiative by the National Academy of Sciences (NAS) to join with peer institutions from around the world to launch international dialogue about policies governing artificial intelligence (AI) and automation. Partners include the National Academy of Engineering, the Canadian National Research Council, the Royal Society, the Royal Academy of Engineering, the Chinese Academy of Sciences, and the Chinese Academy of Engineering. Participants will include government officials, industry leaders, and academic researchers from many different countries in addition to the United States, U.K., China, and Canada. Topics to be addressed include national security, data use and privacy, and legal and intellectual property conundrums related to AI. Grant funds will partially support a workshop and associated webcast, a subsequent workshop report, and the creation and dissemination of supplementary resources for participants and the public.