Homomorphic Encryption for Privacy-Friendly Augmented Democracy

Augmented democracy is a proposal to expand the ability of citizens to participate in the democratic decision process through a digital twin. Artificial intelligence would be used to diminish the load of citizens by recommending decisions based on expert knowledge and the citizens learned preferences. This paper explores the possibility to design citizen's avatars in a privacy preserving way. We formulate the problem as a Collaborative Filtering recommendation system and solve it with Matrix Factorisation. We use Homomorphic Encryption to build two privacy-preserving protocols and evaluate the practicality of our solutions with a toy example using the HEAAN encryption library.