CAREER: Advancing Methods for Analyzing the Effects of Targeted Advertising

Anyone using the internet today is exposed to dozens of targeted digital advertisements online. These advertisements represent an enormous number of different advertisers, and are displayed to internet users in dozens of different ways. As a result, many questions about targeted advertising online arise, including: What and how many ads are people exposed to online? How do these ads differ between different people or groups? How do they change over time in response to people’s online activities? And, most importantly, what effects do they have on people? Addressing these questions can help improve public policy and regulation for the advertising industry, and provide insights and guidance to Americans, whose online experiences are being shaped by targeted advertisements. The research team will build software systems to collect and analyze the content and impacts of digital targeted advertisements on real users. The project also includes initiatives to promote STEM engagement among local high school students in the Philadelphia area. Targeted digital advertisements are both highly personalized and executed through dozens of intermediaries. As a result, understanding the targeting mechanisms, content, and impacts of targeted digital advertising requires an end-user-centered approach that focuses on what content people actually see, and what effect it has on those real people. To do this, the project will make advances in algorithmic auditing methods, which allow people without access to systems’ internal workings to study their outputs and impacts. The research team will develop two systems of auditing infrastructure to enable multi-platform auditing and real-time ad analysis, deploying them with live participants to run experiments that causally measure the impacts of current targeted ads and user-centered alternatives. The project will contribute innovations in sociotechnical AI auditing methods, an open-access Targeted Ad Observatory for analysis by other research teams, and educational materials to engage local STEM students. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.