How Humans Judge Machines


Recent advances in artificial intelligence have spurred a growing debate on how to think about ethics in a world of semi-intelligent machines. This debate, however, has lacked counterfactual measures comparing people's judgment of machines with the judgment of equivalent human actions. In this presentation, I inform this debate by presenting dozens of experiments designed to compare the way in which people judge machines and humans in a variety of settings. The experiments touch on issues of algorithmic bias, privacy, law-enforcement, self-driving vehicles, and labor displacement. I then use statistical methods to model the behavior observed across these experiments, revealing general principles that govern the way in which people judge machines compared to humans. These findings are part of the forthcoming book: How Humans Judge Machines (MIT Press, 2020).

César Hidalgo
Speaker Title
Chair, Artificial and Natural Intelligence
Speaker Institution
Toulouse Institute (ANITI)
Speaker Biography

César A. Hidalgo is a Chilean-Spanish-American physicist, author, and entrepreneur. He currently holds a Chair at the Artificial and Natural Intelligence Toulouse Institute (ANITI), an is also an Honorary Professor at the University of Manchester and a Visiting Professorship at Harvard's School of Engineering and Applied Sciences.  From 2010 to 2019, Hidalgo led MIT’s Collective Learning group and was an Associate Professor at MIT. Prior to joining MIT, Hidalgo was a research fellow at Harvard. Hidalgo is also a founder of Datawheel, an award-winning company specialized in the creation of data distribution and visualization systems. Hidalgo holds a PhD in Physics from the University of Notre Dame and a Bachelor in Physics from Universidad Católica de Chile.