ABSTRACT
The success of platforms such as Facebook and Google has been due in no small part to features that allow advertisers to target ads in a fine-grained manner. However, these features open up the potential for discriminatory advertising when advertisers include or exclude users of protected classes---either directly or indirectly---in a discriminatory fashion. Despite the fact that advertisers are able to compose various targeting features together, the existing mitigations to discriminatory targeting have focused only on individual features; there are concerns that such composition could result in targeting that is more discriminatory than the features individually.
In this paper, we first demonstrate how compositions of individual targeting features can yield discriminatory ad targeting even for Facebook's restricted targeting features for ads in special categories (meant to protect against discriminatory advertising). We then conduct the first study of the potential for discrimination that spans across three major advertising platforms (Facebook, Google, and LinkedIn), showing how the potential for discriminatory advertising is pervasive across these platforms. Our work further points to the need for more careful mitigations to address the issue of discriminatory ad targeting.
Supplemental Material
- 12 CFR § 202.4 (b) - Discouragement. https://www.law.cornell.edu/cfr/text/12/202.4.Google Scholar
- 24 CFR § 100.75 - Discriminatory advertisements, statements and notices. https://www.law.cornell.edu/cfr/text/24/100.75.Google Scholar
- 29 USC § 623 - Prohibition of age discrimination. https://www.law.cornell.edu/uscode/text/29/623.Google Scholar
- M. Ali, P. Sapiezynski, M. Bogen, A. Korolova, A. Mislove, and A. Rieke. Discrimination through Optimization: How Facebook's Ad Delivery can Lead to Biased Outcomes. CSCW, 2019.Google Scholar
- M. Ali, P. Sapiezynski, A. Korolova, A. Mislove, and A. Rieke. Ad Delivery Algorithms: The Hidden Arbiters of Political Messaging. arXiv preprint arXiv:1912.04255, 2019.Google Scholar
- About Audiences for Special Ad Categories. https://www.facebook.com/business/help/2220749868045706.Google Scholar
- About Customer Match. https://support.google.com/adwords/answer/6379332?hl=en.Google Scholar
- D. Biddle. Adverse Impact and Test Validation: A Practitioner's Guide to Valid and Defensible Employment Testing. Gower, 2005.Google Scholar
- S. Barocas and A. D. Selbst. Big data's disparate impact. Cal. Law Rev., 104, 2016.Google Scholar
- A. Datta, A. Datta, J. Makagon, D. K. Mulligan, and M. C. Tschantz. Discrimination in Online Personalization: A Multidisciplinary Inquiry. FAT*, 2018.Google Scholar
- A. Datta, M. C. Tschantz, and A. Datta. Automated Experiments on Ad Privacy Settings: A Tale of Opacity, Choice, and Discrimination. PETS, 2015.Google Scholar
- E. Dreyfuss. Facebook Changes Its Ad Tech to Stop Discrimination. WIRED. https://www.wired.com/story/facebook-advertising-discrimination-settlement/.Google Scholar
- L. Edelson, S. Sakhuja, R. Dey, and D. McCoy. An Analysis of United States Online Political Advertising Transparency. https://arxiv.org/abs/1902.04385.Google Scholar
- Exhibit a - Programmatic Relief. https://nationalfairhousing.org/wp-content/uploads/2019/03/FINAL-Exhibit-A-3-18.pdf.Google Scholar
- Facebook Engages in Housing Discrimination with Its Ad Practices, U.S. Says. https://www.nytimes.com/2019/03/28/us/politics/facebook-housing-discrimination.html.Google Scholar
- Facebook Lets Advertisers Exclude Users by Race. https://www.propublica.org/article/facebook-lets-advertisers-exclude-users-by-race/.Google Scholar
- Facebook: About Facebook Pixel. https://www.facebook.com/business/help/742478679120153.Google Scholar
- Facebook: About Lookalike Audiences. https://www.facebook.com/business/help/164749007013531.Google Scholar
- Facebook: Best Practices for Detailed Targeting. https://www.facebook.com/business/help/905095143159925.Google Scholar
- Facebook: Brand Safety Controls. https://www.facebook.com/business/help/1926878614264962.Google Scholar
- Facebook: Use Detailed Targeting. https://web.archive.org/web/20200429194031/https://www.facebook.com/business/help/440167386536513.Google Scholar
- Get Started with LinkedIn Website Retargeting. http://web.archive.org/web/20200601191411/https://www.linkedin.com/help/lms/answer/73934.Google Scholar
- Google: About audience targeting. https://support.google.com/google-ads/answer/2497941?hl=en.Google Scholar
- Google: About contextual targeting. https://support.google.com/google-ads/answer/2404186?hl=en.Google Scholar
- Google: About managed placements. https://support.google.com/google-ads/answer/2470108?hl=en.Google Scholar
- Google: About remarketing setup. https://support.google.com/google-ads/answer/2454000?hl=en.Google Scholar
- Google: About similar audiences on the Display Network. https://support.google.com/google-ads/answer/2676774?hl=en.Google Scholar
- Google: About topic targeting. https://support.google.com/google-ads/answer/2497832?hl=en.Google Scholar
- Google: Frequency capping: Definition. https://support.google.com/google-ads/answer/117579?hl=en.Google Scholar
- Google: How to use combined audiences. https://support.google.com/google-ads/answer/9066030?hl=en.Google Scholar
- HUD Sues Facebook Over Housing Discrimination and Says the Company's Algorithms Have Made the Problem Worse. https://www.propublica.org/article/hud-sues-facebook-housing-discrimination-advertising-algorithms.Google Scholar
- A. Lambrecht and C. E. Tucker. Algorithmic bias? An empirical study into apparent gender-based discrimination in the display of STEM career ads. 2018. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2852260.Google Scholar
- LinkedIn Audience Network - Manage Delivery Preferences. https://www.linkedin.com/help/linkedin/answer/83630.Google Scholar
- LinkedIn Contact Targeting. https://business.linkedin.com/marketing-solutions/ad-targeting/contact-targeting.Google Scholar
- LinkedIn: Use and-OR Targeting for Objective-Based Advertising Campaigns. https://www.linkedin.com/help/lms/answer/102905/use-and-or-targeting-for-objective-based-advertising-campaigns?lang=en.Google Scholar
- L. Sweeney. Discrimination in online ad delivery. CACM, 56(5), 2013.Google Scholar
- T. Speicher, M. Ali, G. Venkatadri, F. N. Ribeiro, G. Arvanitakis, F. Benevenuto, K. P. Gummadi, P. Loiseau, and A. Mislove. On the Potential for Discrimination in Online Targeted Advertising. FAT*, 2018.Google Scholar
- Targeting Options for LinkedIn Advertisements. https://web.archive.org/web/20200525174210/https://www.linkedin.com/help/lms/answer/722/targeting-options-for-linkedin-advertisements?lang=en.Google Scholar
- Targeting with LinkedIn Lookalike Audiences - Overview. https://www.linkedin.com/help/lms/answer/94287.Google Scholar
- Upturn Amicus Brief in Onuoha v. Facebook. https://www.courtlistener.com/recap/gov.uscourts.cand.304918/gov.uscourts.cand.304918.76.1.pdf.Google Scholar
- G. Venkatadri, Y. Liu, A. Andreou, O. Goga, P. Loiseau, A. Mislove, and K. P. Gummadi. Privacy Risks with Facebook's PII-based Targeting: Auditing a Data Broker's Advertising Interface. IEEE S&P, 2018.Google Scholar
- What's a Custom Audience from a Customer List? https://www.facebook.com/business/help/341425252616329/.Google Scholar
Index Terms
- On the Potential for Discrimination via Composition
Recommendations
Is Combining Contextual and Behavioral Targeting Strategies Effective in Online Advertising?
Online targeting has been increasingly used to deliver ads to consumers. But discovering how to target the most valuable web visitors and generate a high response rate is still a challenge for advertising intermediaries and advertisers. The purpose of ...
An economic analysis of online advertising using behavioral targeting
Online publishers and advertisers have recently shown increasing interest in using targeted advertising online. Such targeting allows them to present users with advertisements that are a better match, based on their past browsing and search behavior and ...
Smart Targeting: A Relevance-driven and Configurable Targeting Framework for Advertising System
RecSys '20: Proceedings of the 14th ACM Conference on Recommender SystemsTargeting system is an essential part of computational advertising. It allows advertisers to select and reach their targeted users. Due to various advertising goals and the demand for making budget plans, advertisers have a strong will to configure the ...
Comments