Bende gij een Brabander is a design exploration that serves to showcase the ability of AI as a facilitator of personalized experiences in a public context. The specific use case provided was that of designing for the Eindhoven-based “Museum door de Stad”, whose ambition it is to entertain visitors with a live interpretation of personal data through AI, in the context of celebrating a hundred years of Eindhoven. The proposed design is an installation which focuses on the pronunciation of a specific sentence in the Brabant dialect: “Bende gij een Brabander?”.

We coded two machine learning models (Gradient Boosting and Random Forest) and one deep learning model (1D Convolutional Neural Network) that we trained on speech samples from both non-Brabant and Brabant speakers. The model was finally able to give a probability of a person having a Brabant accent. This factor is then used to provoke discussion on what it means to be a Brabander, by means of theses, with which a participant can agree or disagree by pinning a printout of their sample to a wall. Through these discussions, we engage participants to look more critically at what it means to be a Brabander, and how speech and accents factor into this identity.

Download project report