Working Papers

This working paper asks organizations to rethink a consequential design challenge: How can the rise of agentic AI strengthen rather than diminish human judgment?

Leif Sigerson (Pinterest), Tom Cunningham (METR), Winston Chou (Netflix), Sana Pandey (MIT CSAIL), Jonathan Stray (UC Berkeley CHAI), Lo-Hua Yuan (Airbnb ), Eytan Bakshy (Meta), Timothy Chan (Statsig), Molly Davies (Pinterest), Maria Dimakopoulou (Uber), Simon Ejdemyr (Netflix), Kenneth Hung (Meta), Nathan Kallus (Netflix & Cornell University), Thu Le (Lyft), A. Demetri Pananos (Datadog), Lee Richardson (Google), Brennan Schaffner (Knight-Georgetown Institute), Rose Tan, Martin Tingley, Nadia Tomova (Booking.com), Panagiotis Toulis (University of Chicago Booth School of Business), Wenjing Zheng (Roblox), Zander Arnao (Knight-Georgetown Institute), Dean Eckles (MIT)

This paper collects and shares industry knowledge on how to make decisions from short-term experiments that are better aligned with long-term outcomes. Based on a daylong workshop with 26 experts from 15 online platforms and 4 universities, it formulates a series of propositions that reflect current industry knowledge.

Useful social science theories predict behavior across settings. However, applying a theory to make predictions in new settings is challenging: rarely can it be done without ad hoc modifications to account for setting-specific factors. We argue that AI agents put in simulations of those novel settings offer an alternative for applying theory, requiring minimal or no modifications.

Abigail Sussman

Yujin Yang

This paper finds that people rely on a competence-judgment tradeoff to choose between human and AI advisors, evaluating whether the benefit of expertise outweighs the cost of social judgment.

Maja S. Svanberg
Brian C. Goehring

This paper’s findings suggest that AI job displacement will be substantial, but also gradual – and therefore there is room for policy and retraining to mitigate unemployment impacts.

Natalia Fischl-Lanzoni

Do leading LLM developers possess a proprietary “secret sauce,” or is LLM performance driven by scaling up compute? Using training and benchmark data for 809 models released between 2022 and 2025, the authors estimate scaling-law regressions with release-date and developer fixed effects.

Authors develop a dynamic model in which a decision-maker chooses AI usage intensity for a worker over time, trading immediate productivity against the erosion of worker skill, testing when decision-makers opt for AI use and the productivity of workers as a result.

John Fallon | Boston University

This paper explores whether human intermediaries add value when employers already have access to algorithmic screening tools.