Should we agree to disagree about Twitter's bot problem?

Official URL: http://dx.doi.org/10.1016/j.osnem.2023.100263

Bots, simply defined as accounts controlled by automation, can be used as a weapon for online manipulation and pose a threat to the health of platforms. Researchers have studied online platforms to detect, estimate, and characterize bot accounts. Concerns about the prevalence of bots were raised following Elon Musk's bid to acquire Twitter. In this work, we want to stress that crucial questions need to be answered in order to make a proper estimation and compare different methodologies and definitions based on behaviors and activities; otherwise the real questions concerning the health of online platforms will be confounded by disagreements about definitions and models. We argue how assumptions on bot-likely behavior, the detection approach, and the population inspected can affect the estimation of the percentage of bots on Twitter. Finally, we emphasize the responsibility of platforms to be vigilant, transparent, and unbiased in dealing with threats that may affect their users.