Skip to content
Technical University of Munich

Wahl.chat: student project supports political opinion-forming

Studies |

An AI-supported tool that enables a new way of engaging with the parties' positions on the 2025 Bundestag elections: a team of students from the TUM School of Computation, Information and Technology, LMU Munich and the University of Cambridge has developed wahl.chat.
The wahl.chat team: Sebastian Maier, Anton Wyrowski, Michel Schimpf, Robin Frasch and Roman Mayr
Behind wahl.chat are the students Sebastian Maier, Anton Wyrowski, Michel Schimpf, Robin Frasch and Roman Mayr.

Wahl.chat is an interactive assistant that answers questions based on the parties' election programs. The tool offers users the opportunity to obtain detailed and source-based information on the parties' positions. The aim is to help voters make informed decisions for the election.

While conventional tools such as the Wahl-O-Mat evaluate predefined theses, wahl.chat allows for open dialog. This enables anyone interested to obtain in-depth information on topics relevant to them and clarify individual questions. The aim of wahl.chat is to make it easier for voters to engage with political issues in an in-depth and up-to-date manner.

How did project wahl.chat come about?

The idea for wahl.chat was born in November 2024 during an informal discussion in Cambridge. Inspired by the desire to enhance existing digital voting aids, the approach of creating an interactive AI tool was developed. The ultimate goal was to give people interested in politics the opportunity to engage with the parties' content in a targeted and personalized way.

The project is backed by a team of five students:

  • Sebastian Maier, student of psychology (LMU Munich)
  • Anton Wyrowski, student of computer science (TUM)
  • Michel Schimpf, PhD candidate (University of Cambridge)
  • Robin Frasch, student of computer science (TUM)
  • Roman Mayr, student of computer science (TUM)

The team combines expertise from different disciplines and works together on a research group at the University of Cambridge on the topic of artificial intelligence.

How does wahl.chat work?

wahl.chat is based on the election programs and other relevant documents of the parties such as the party manifestos. The GPT-4o language model from OpenAI is used to generate answers to user questions. These answers are based on relevant text excerpts and are designed to be fact-based, neutral and transparent. Additional sources, such as news portals, are included via the Perplexity.ai service for a differentiated classification of the answers. Thereby, the underlying Large Language Model (LLM) focuses on high-quality sources.

Special attention is paid to the security and protection of user data: wahl.chat does not store any personal data and works completely anonymously.

Further information about the project and the team can be found at: wahl.chat.

Links:

◄ Back to: News & Events
To top