About Us
We are an interdisciplinary team dedicated to bringing transparency to the use of AI in political discourse on social media. Our project has received significant support through multiple funding sources, enabling us to build a comprehensive monitoring infrastructure.
Funding & Support
Our work is made possible through grants from the University of Amsterdam’s Research Priority Area in Artificial Intelligence & Politics (AI & Politics RPA), which provides annual funding of 250,000 euros over five years to stimulate interdisciplinary research into how AI is reshaping politics and society.
We also received crucial funding from the SIDN Fonds as part of the Hybrid Election Interference Observatory (HEIO) consortium. This project brings together Post-X Society, AI Forensics, Trollrensics, and Justice for Prosperity to protect election integrity by monitoring social media platforms for foreign interference operations.
Our Team
The CampAIgn Tracker is led by Dr. Fabio Votta (AlgoSoc Consortium at the University of Amsterdam) and Dr. Simon Kruschinski (University of Mainz), who brought this initiative to life and are responsible for the research project. Both researchers specialize in political campaigns on social media.
Research Assistants
Doortje Cools, Lynn Louvet, Ella van Vloten, Daisy Blass, Isabella van der Zanden, Emily Murphy, Raluca Groza
We extend our deepest thanks to our research assistants for their exceptional dedication and effort in bringing this project to life.
PhD Students & Postdocs (University of Amsterdam)
Tobias Kleindamm, Ellen Linnert, Annamijn Beijeman, Haylee Kelsall,
We are grateful to our colleagues, whose expertise and commitment have been vital for the success of this project.
Method
With the CampAIgn Tracker, we aim to help citizens and journalists understand where AI has been used and in what context. To achieve this, we have developed a two-stage process to reliably identify and code AI-generated images.
Data Collection
We capture all posts from over 4000 accounts of parties, leading candidates, political commentators and influencers. We consider party accounts at the national, provincial, and local levels.
We are currently tracking 5463 posts between 2025-09-17 and 2025-09-25.
The classification of accounts is based on the work of our research assistants as well as dedicated volunteer work of Who Targets Me, an organization committed to transparency in digital campaigning. They track the online advertising spending of over 100,000 advertisers from 650+ parties in over 50 countries.
AI Detection
Our AI detection is a two-stage process that combines automatic classification using an AI model with manual annotation.
1. Automatic Review: - All collected images and video stills are checked for AI generation using the AI detection model from SightEngine.
2. Human Validation: - If the classification probability is above 0.1, trained research assisstants manually review the image. - This ensures that automatic AI classification is always validated by human expertise.
Additionally, for all identified AI posts, coders classify the following features:
- The labeling of AI use
- The actors in the image or video
- The theme of the image or video
- The actors in the entire post
- The use of negativity, e.g., through attacks on parties
- The use of acclaims, e.g., through references to a party’s successes
This comprehensive information is available on our platform campaigntracker.nl in simple visualizations.
Our hope is that this enables citizens and journalists to learn how AI is being used in election campaigns and to hold a public debate about the use of AI in politics.
Our Origins
We initially started with this project at the Politechathon in Berlin, organized by the Baden-Württemberg Stiftung, where we won the prize in the “Media Literacy” category.
The CampAIgn Tracker project launched its first version during the 2025 German elections: campaigntracker.de.
From left to right: Theresa Gröninger (Jury member at the Politechathon), Theresa Schültken, Miriam Runde, Fabio Votta, Simon Kruschinski, Jakob Scherer
[Credits: @BW Stiftung / Viktor Heekeren]
