How Community Notes Feature in Twitter is Used in Japan

Abstract

In recent years, X (formerly Twitter) introduced a feature known as "Community Notes" (formerly Birdwatch), a system where users can propose annotations to posts that contain misinformation. If these annotations are deemed helpful by a certain evaluation, they are then displayed additionally. Unlike being reviewed by professional institutions, this mechanism entrusts the judgement to a multitude of users, anticipating a more rapid correction of misinformation. However, concerns have been raised about the potential lack of sufficient quality assurance and the possibility of partisanship affecting evaluations. This paper demonstrates the existence of behaviors, such as attaching numerous notes to specific users, and through case analysis via clustering of notes written in Japanese, it elucidates the challenges of crowdsourced fact-checking.