AI, ads, and attention: The role of algorithms in political content exposure

Partners: University of Amsterdam 
Type: Postdoc
Duration: 2023—2025

In the contemporary digital landscape, algorithmic systems are pivotal in shaping what political information citizens are exposed to. This postdoctoral research project, in collaboration with the University of Amsterdam and the AlgoSoc program, aims to develop robust research infrastructures that enhance transparency in digital political advertising and digital news engagement. By employing innovative methods such as data donations and computational analyses, the project seeks to illuminate the inner workings of (ad delivery algorithms) and their impact on public discourse, particularly during election periods.

A central component of this research involves the creation of digital infrastructures that facilitate the collection and analysis of data related to political advertisements and news consumption. Utilizing data donation frameworks, participants voluntarily share their digital trace data, enabling researchers to gain unprecedented insights into the targeting strategies and dissemination patterns employed by political actors.

Key Research Questions:

1. What role do user behaviors (e.g., page likes, interactions) and network characteristics (e.g., political homophily) play in algorithmic classification?

2. To what extent do Meta’s ad delivery algorithms influence who sees political content, and at what cost?

3. How does audience segmentation affect ideological diversity in digital news and ad exposure?

4. Are certain demographic groups systematically excluded or miscategorized in inferred political interest classifications?

5. Do people perceive themselves as less influenced by targeted political ads compared to others, and does this vary across different audience segments?