AI Cyber Lunch: Piotr Sapiezynski on "Discrimination Through Optimization"

Name: AI Cyber Lunch: Piotr Sapiezynski on "Discrimination Through Optimization"
Start: 2023-03-01T12:00:00-0500
End: 2023-03-01T13:00:00-0500

Open to the Public

Please join the Science, Technology, and Public Policy Program for an AI Cyber Lunch Seminar featuring Piotr Sapiezynski, Associate Research Scientist at the Khoury College of Computer Sciences at Northeastern University. In a talk entitled "Discrimination Through Optimization," Sapiezynski will discuss the role of online advertising platforms' delivery optimization processes in perpetuating - and potentially preventing - discrimination in digital advertising.

Q&A to follow. Buffet-style lunch will be served.

Registration: In-person attendance is limited to current Harvard ID holders. No RSVP is required. Room capacity is limited and seating will be on a first come, first served basis.

Members of the public are welcome to attend virtually via Zoom. Virtual attendees should register using the button below; upon registering, attendees will receive a confirmation email with a Zoom link.

Recording: Please be advised that this seminar will not be recorded.

Accessibility: Persons with disabilities who wish to request accommodations or who have questions about access, please contact Liz Hanlon (ehanlon@hks.harvard.edu) in advance of the session.

Date: Wed, Mar 1, 2023 Time: 12:00pm - 01:00pm Location: Room 434 A-B

Summary

The enormous financial success of online advertising platforms is partially due to the precise targeting features they offer. Although researchers and journalists have found many ways that advertisers can target—or exclude—particular groups of users seeing their ads, comparatively little attention has been paid to the implications of the platform’s ad delivery process, comprised of the platform’s choices about which users see which ads. It has been hypothesized that this process can “skew” ad delivery in ways that the advertisers do not intend, making some users less likely than others to see particular ads based on their demographic characteristics. In this talk I will demonstrate that such skewed delivery occurs on Facebook, due to market and financial optimization effects as well as the platform’s own predictions about the “relevance” of ads to different groups of users. We observe significant skew in delivery along gender and racial lines for “real” ads for employment and housing opportunities despite neutral targeting parameters. We also measure skews in delivery of political ads along inferred political leaning. Our results demonstrate mechanisms that can lead to potentially discriminatory ad delivery, even when advertisers set their targeting parameters to be highly inclusive. This underscores the need for policymakers and platforms to carefully consider the role of the ad delivery optimization run by ad platforms themselves—and not just the targeting choices of advertisers—in preventing discrimination in digital advertising.

Speaker

Piotr Sapiezynski is an Associate Research Scientist at the Khoury College of Computer Sciences at Northeastern University in Boston, MA. He received a PhD in Network Science/Data Science from Technical University of Denmark. The core of his work is auditing platforms and their algorithms for fairness and privacy. He investigates systems that are optimized for corporate profit yet drive many aspects of our daily lives. All too often he find these systems have (possibly unintended but often predictable) side effects that bring harm to individuals and the society.

His findings on Facebook's use of personal data without consent were a part of a record $5B dollar settlement with the US Federal Trade Commission. The Department of Justice settled with Facebook over issues documented through his research, in the first ever case where the federal government challenged algorithmic discrimination under the Fair Housing Act. He briefed Members of the House Financial Services Committee on discriminatory effects in ad delivery. He presented at a public hearing to the European Parliament's Internal Market and Consumer Protection Committee on price differentiation and negative effects of personalization in political advertising. He was a member of the Public Health, Surveillance, and Human Rights Network in the wake of the COVID-19 pandemic.