Being an Upstander: Overcoming Hate Speech in Social Media

There has been an increase in hate speech, most of it motivated by ethnic or religious discrimination. In this sense, social media has been one of the channels used for its propagation. Despite this, scientific literature has paid much less attention to the discourse that tries to stop it with the counter-narrative. Thus, from the European projects ALRECO and REALUP (CERV 2018-2021/ 2021-2024) and within the framework of the X network, different intensities of hate speech have been identified, including in the analysis the presence of Upstander Speech (referring to the narrative that responds to or anticipates hate speech). At first, an algorithm was designed, and a machine learning process was followed to automatically identify X messages by intensities in hate speech, the existing alternative and affected social groups. In a second part, it was relevant to identify the most effective strategies of Upstander speech. Thus, the results indicate that strengthening the online (and offline) community is a key element to counteract the effects of ‘isolating violence’ and thus a greater positioning of users against hate speech. On the other hand, the analysis through Social Media Analytics (SMA) has revealed that the messages with the greatest impact (with more Retweets and more likes) are those that contain scientific evidence and use attractive and non-violent language. The analysis reveals some strategies and actions that increase the effectiveness and number of Upstander messages, thus favoring greater visibility of Upstander speech. This has a positive impact on the reduction of hate speech by offering information to those followers of hate who show doubts, but also by not normalizing hate speech it invites upstander positioning in favor of the victims.