The Adversarial Collaboration Project (ACP) is a research initiative at University of Pennsylvania. Its goal is to improve the accuracy of the social and behavioral sciences, and they seek to achieve this by popularizing adversarial collaborations as the approach to resolving scientific disputes. They aim to discover best practices for participating in adversarial collaborations, making them more effective, and to normalize such practices in order to improve the accuracy and efficiency of research in the social sciences and its reputation among policymakers and the public.

ACP seeks to popularize this scientific approach by supporting a series of high impact inaugural projects, engaging many high-profile scholars and scholarly disputes. The projects are wide ranging and cover questions in the social sciences such as: “Is reasoning socially motivated?”; “Do accuracy nudges reduce misinformation sharing across the ideological spectrum?”; and “Is there systematic censorship in social science?” The goal is for these projects to be some of the highest quality and most talked-about projects to be published in the social sciences this decade, both within the social sciences and in traditional popular and social media. 

The project tackles many intertwined challenges simultaneously: advancing scientific debates, elevating pursuit of truth in science, reducing hostility and polarization in science, promoting evidence-based policy-making, and increasing public trust in science. Ultimately, the project aims to demonstrate to scientific peers that adversarial collaborations are feasible, rewarding, and essential for scientific progress.