24th Conference of the International Federation of Operational Research Societies
Abstract Submission

3372. A Model of AI Supply Chain Governance for Social Media Addiction

Invited abstract in session HF-14: OR conjectures for AI, stream OR and Ethics, and Societal Perspectives.

Thursday, 16:15-17:45
Room: HG – Hörsaal 01

Authors (first author is the speaker)

1. James Siderius
Tuck School of Business, Dartmouth College
2. Laurens Debo
Dartmouth College

Abstract

Social media platforms use compute to improve recommendations by learning about content and users. While compute can surface valuable content, it can also enable platforms to target vulnerable users with addictive material. We develop a model where a platform chooses its interface design and allocates compute between learning about content or users, while regulators intervene through compute caps, algorithmic constraints, or litigation. We show that compute caps can have non-monotonic welfare effects, with the effectiveness of different policies depending on the degree of behavioral targeting.

Keywords

Status: accepted


Back to the list of papers