Detection Practices of Hate Speech: A University e-Tivity Case Study

1 Hate speech: “ambiguous but useful” category Contemporary forms of hatred (hate speech) encompass a broad spectrum of behaviours, yet common features unite them, particularly concerning the social and psychological dynamics that drive human hostility, which can vary individually. Understanding these dynamics is essential for a non-naïve or overly idealistic education that addresses and challenges these phenomena within their cultural, social, and political contexts [1]. In the social web various forms of hate speech and flame wars (flaming) are experienced, amplified, and sometimes countered. These flame wars are marked by the violent polarization of viewpoints, where personal attacks overshadow the content of debates, diverting attention from the issue at hand [2]. Within the field of Hate Studies [3], the definition and boundaries of "hate speech" remain contested, as various disciplinary approaches often diverge in their interpretations. It is an “ambiguous but useful category” [4]. This ambiguity also emerges in research employing algorithmic automation for hate speech detection [5]. A useful definitional framework is CM/Rec(2022)16 by the Committee of Ministers of the Council of Europe (132nd session), which categorises hate speech into three distinct categories, each requiring different responses and measures. The third level concerns cases that fall outside the scope of criminal or administrative liability, yet “can nevertheless cause or amplify prejudice, intolerance and hatred, raise concerns in terms of tolerance, civility, inclusion and respect for the rights of others, and threaten social cohesion and democratic stability” [6]. The decision to include this third level, which we might refer to as the ‘citizenship’ level, in the definition of the phenomenon has important implications for countering online hate and call for a media and educational approach [7]. Accordingly, higher education will be examined as a pedagogical space in which, consistent with the disciplinary orientations of academic programmes, this kind of sensitivity can be meaningfully cultivated. 2 Case Study: An e-Tivity Designed within an Information Literacy Course The case study presented focuses on the e-tivity conducted during the academic years 2023–2024 and 2024–2025 within the Information Literacy course of the Master's Degree in Media Education (LM-93) at the Catholic University of the Sacred Heart in Milan. Students were asked to carry out a research and recognition task focused on identifying cases online hate speech on TikTok, within the broader polarised discursive landscape of the current conflict in Palestine (diverse groups targeted). In addition to a first-level evaluation of whether the content could be considered hateful or not, students were required to conduct a media analysis of TikTok videos and subsequently apply the “Online Hate Spectrum” – an interpretive framework developed through research conducted by the Mediavox Observatory at the Catholic University [8]. The purpose of this interpretive framework was to overcome the simplistic binary of “hate” versus “no hate,” fostering a more nuanced understanding of online hostility. Over the two academic years, the e-tivities led students to collect and analyse a total of 408 TikTok videos. The cases were examined using interpretive categories of hate speech studied during the course [9]. This process enabled students to critically engage with key phenomena such as the role of fast thinking in heuristics [10], the emergence of new authorship standards on social media, the banalisation of content in flame wars, the rhetoric linked to anonymity, the impact of the spiral of silence on social platforms, digital illiteracy, the cultural role of memes, the post-truth framework, and the concept of margin effect [11]. The most controversial cases were selected for in-class discussion, encouraging peer debate and critical reflection among students. 3 Conclusion The e-tivity thus frames the university course as an opportunity to apply knowledge on online hate within a specific media context—namely, the discursive flows on TikTok related to the conflict in Palestine. More specifically, the individual-level detection activity followed by collective classroom discussion fostered media-educational awareness regarding the interpretation of online hate. This experience confirmed the pedagogical value of an interpretive tool such as the Online Hate Spectrum, precisely because it offers a more nuanced and multifaceted approach compared to a simplistic binary logic of yes/no.