ABSTRACT
NewsButler will explore new channels of personalized news distribution and interaction based on conversational user interfaces driven by recommendations that are based on machine learning algorithms, reader configuration, and editorial validation. Today’s news consumers appreciate personalized content recommendations but are critical of fully algorithm-based personalization, which has the tendency to lead to a filter bubble.

“NewsButler will create a smart content stream recommender and editorial optimalization engine combining human and machine intelligence that helps readers and editorial teams create a personalized, user controlled news experience.”

The project will research and develop a demonstrator of an intelligent ‘content stream’ recommender and editorial optimization engine that, like a digital butler, assists both readers and editorial teams to serve every single reader a personal, user-controlled news experience completely tailored to his interests. The NewsButler consortium is convinced that the combination of intelligent recommendations within innovative bot-based distribution channels, tested among a large panel of users, will lead to new kinds of monetization models for the publishing industry. To achieve these innovation goals, the project involves collaboration between research and industry partners active in machine learning, news production, business and user modelling, and user experience design.