Who is paying your EU Commissioner?

For ETH Zürich’s “Hacking Democracy” hackathon, our team explored how emerging technologies could help journalists and researchers investigate lobbying. We built a prototype that maps the EU Transparency Register and other publicly available databases into a knowledge graph, helping users follow connections between lobbying organisations, politicians, and the laws they pass.

A swarm of language-model agents helps organise source material and populate the Neo4j graph. Journalists, researchers, and agentic AI systems can then query those structured relationships directly instead of asking a model to reread thousands of documents for every question. This approach is intended to make investigation faster, reduce unsupported connections, and use less energy. Its accuracy and energy use still need systematic verification, as does the process of adding further datasets.

“Hacking Democracy” brings students from different disciplines together to prototype open-source tools for civic engagement, deliberation, and transparency. The hackathon introduces participatory methods such as citizens’ assemblies and participatory budgeting, then asks teams to apply them to a challenge related to participatory democracy.

Through iterative design, development, and testing, teams explore how technology can support more inclusive and effective collective decision-making. EU-Daemon was our team’s prototype for investigating transparency and lobbying data.