Privacy and Security Threats in Social Network Ad Targeting and Delivery

The enormous financial success of online advertising platforms is partially due to their precise targeting and delivery features. Recently, such platforms have been criticized for allowing advertisers to exclude users belonging to a certain race or gender from receiving their ads. In this Cyber Alliance talk, Northeastern Prof. Alan Mislove will discuss two threads of work that aim to understand the extent of this discrimination.

First, he will examine the advertisers' choices about who they wish to bid on. Most platforms allow advertisers to target users directly by uploading their personally identifiable information (PII), but it remains unclear which sources of PII ad platforms use. Focusing on Facebook, Prof. Mislove will demonstrate that despite all privacy settings being enabled, phone numbers and email addresses added for security purposes (e.g., two-factor authentication), those provided to the Facebook Messenger app, and those included in friends’ uploaded contact databases are all used by Facebook to allow advertisers to target users.

Second, he will examine the platform's choices about who should see an ad. Prof. Mislove will demonstrate that ad delivery can be significantly skewed on Facebook, due to the platform's predictions about the "relevance" of ads to different groups. He will demonstrate significant skew in delivery along gender and racial lines for "real" ads for employment and housing opportunities. His findings have detailed previously unknown mechanisms that could lead to discriminatory ad delivery, even when advertisers set their targeting parameters to be highly inclusive.

There will be time for casual conversation and light refreshments before and after the presentation. Please RSVP to tgabs@bu.edu.