Building Human Values into Recommender Systems: An Interdisciplinary Synthesis
26 Sep 2022
A massive industry-academic-civil society collaborative paper published for arXiv, led by CHAI’s Jonathan Stray, and involving Dylan Hadfield-Menell from CHAI along with Alon Halevy, Parisa Assar, Craig Boutilier, Amar Ashar, Lex Beattie, Michael Ekstrand, Claire Leibowicz, Connie Moon Sehat, Sara Johansen, Lianne Kerlin, David Vickrey, Spandana Singh, Sanne Vrijenhoek, Amy Zhang, McKane Andrus, Natali Helberger, Polina Proutskova, Tanushree Mitra and Nina Vasan.
The paper catalogues the values that seem most relevant to AI-driven content personalization algorithms. It explains how existing systems work, design approaches for implementing various values, and gives policy options for regulating them.