The problem of creepy personalised recommendations in banks

Abstract

This thesis set out to investigate two primary questions: (1) what makes some personalised recommendations from banks feel creepy, and (2) how banks can maintain highly personalised recommendations while eliminating creepiness. To address these questions, the thesis employed a strategy of first establishing a theoretical framework, examining cases where the risk of creepiness is lower or absent compared to the same recommendation given in different context, and, finally, proposing some practical solutions for banks based on the insights gained, aiming to preserve the benefits of personalisation and eliminate the harms of creepiness.