“Get lost, GetLostBot!” annoying people by offering recommendations when they are not wanted

摘要

This brief paper outlines the experience of releasing a purposefully contrary recommendation service for Foursquare called GetLostBot.udGetLostBot works differently than most recommender systems since it is responsive to behaviours rather than user requests. The system automatically monitors Foursquare behaviour and intervenes with suggestions when users fall into a routine. These interventions take the form of mysterious walking directions that challenge the user to visit somewhere new. Importantly, these suggestions are explicitly not informed by traditional metrics such as popularity, high ratings, or friend activity, and instead act as prompts to explore unknown places.udThis paper discusses the reception to the application, highlighting the apparent disconnect between users’ good intentions around becoming more serendipitous, and the reality of those interventions as they are experienced in the wild.