Accountability Issues, Online Covert Hate Speech, and the Efficacy of Counter‐Speech
Open accessAbstract
Concerning individual or institutional accountability for online hate speech, research has revealed that most such speech is covert (veiled or camouflaged expressions of hate) and cannot be addressed with existing measures (e.g., deletion of messages, prosecution of the perpetrator). Therefore, in this article, we examine another way to respond to and possibly deflect hate speech: counter-speech. Counter-narratives aim to influence those who write hate speech, to encourage them to rethink their message, and to offer to all who read hate speech a critical deconstruction of it. We created a unique set of parameters to analyze the strategies used in counter-speech and their impact. Upon analysis of our database (manual annotations of 15,000 Twitter and YouTube comments), we identified the rhetoric most used in counter-speech, the general impact of the various counter-narrative strategies, and their specific impact concerning several topics. The impact was defined by noting the number of answers triggered by the comment and the tone of the answers (negative, positive, or neutral). Our data reveal an overwhelming use of argumentative strategies in counter-speech, most involving reasoning, history, statistics, and examples. However, most of these argumentative strategies are written in a hostile tone and most dialogues triggered are negative. We also found that affective strategies (based on displaying positive emotions, for instance) led to a positive outcome, although in most cases these narratives do not receive responses. We recommend that education or training—even machine learning such as empathetic bots—should focus on strategies that are positive in tone, acknowledging grievances especially.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Affective rhetoric has been shown to be effective.
“since affective rhetoric has been shown to be effective (Baider, 2023), we focus on two emotions”
From “Real men in the battlefield”: a hybrid AI–discourse analysis of online homophobia and transphobia (2026) · cites this paper as Baider, 2023
The Abstract reports positive outcomes for affective strategies that display positive emotions, and §4.2.3 links acknowledging grievances with positive responses. But §4.3 says negative affective rhetoric exacerbated verbal violence, so the paper does not support the unqualified claim that affective rhetoric is effective. (LLM, conf=0.88)
Experimental — accuracy not yet measuredThe annotation schema incorporated both argumentative and affective strategies and was based on Baider (2023) for emotions and Baider (2025) for the frames.
“The annotation schema we used incorporated both argumentative and affective strategies. It was based on Baider (2023) for emotions, and Baider (2025) for the frames.”
From “Real men in the battlefield”: a hybrid AI–discourse analysis of online homophobia and transphobia (2026) · cites this paper as Baider (2023)
The negativity of the counter speech was expected.
“The negativity of the counter speech was expected (Baider, 2023).”
From “Real men in the battlefield”: a hybrid AI–discourse analysis of online homophobia and transphobia (2026) · cites this paper as Baider, 2023
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 36 · Cited by 33
Research works printed but not linked: 29
- Ashida, M., & Komachi, M. (2022). Towards auto- matic generation of messages countering online hate speech and microaggressions. In A. Narang, A. Davani, L. Mathias, B. Vidgen, & Z. Talat (Eds.), Pro- ceedings of the sixth workshop on online abuse and harms (pp. 11-23). Association for Computational Linguistics.
- Askanius, T. (2021). On frogs, monkeys, and execution memes: Exploring the humor-hate nexus at the inter- section of neo-nazi and alt-right movements in Swe- den. Television & New Media, 22(2), 147-165.
- Baider, F. (2019). Le discours de haine dissimulée; le mépris pour humilier [Covert hate speech; using contempt to humiliate]. Déviance et Société, 43(1), 71-100.
- Baider, F. (2020). Pragmatics lost? Overview, synthesis and proposition in defining online hate speech. Prag- matics and Society, 11(2), 196-218.
- Baider, F. (2022). Covert hate speech, conspiracy the- ory and anti-semitism: Linguistic analysis versus legal judgement. International Journal for the Semiotics of Law, 35, 1-25. https://doi.org/10.1007/s11196-022- 09882-w · DOI
- Ben-David, A., & Fernández, A. M. (2016). Hate speech and covert discrimination on social media: Monitor- ing the Facebook pages of extreme-right political par- ties in Spain. International Journal of Communication, 10, 1167-1193.
- Bhat, P., & Klein, O. (2020). Covert hate speech: White nationalists and dog whistle communication on Twit- ter. In G. Bouvier & J. Rosenbaum (Eds.), Twitter, the public sphere, and the chaos of online deliberation. Palgrave Macmillan. https://doi.org/10.1007/978-3- 030-41421-4_7 · DOI
- Braddock, K., & Horgan, J. (2016). Towards a guide for constructing and disseminating counternarratives to reduce support for terrorism. Studies in Conflict & Ter- rorism, 39(5), 381-404.
- Briggs, R., & Feve, S. (2013). Review of programs to counter narratives of violent extremism. What works and what are the implications for govern- ment? Institute for Strategic Dialogue. https://www. publicsafety.gc.ca/lbrr/archives/cn28580-eng.pdf
- Burnap, P., & Matthew, L. W. (2015). Cyber hate speech on Twitter: An application of machine classification and statistical modeling for policy and decision mak- ing. Policy Internet, 7(2), 223-242.
- Citron, D. K., & Norton, H. (2011). Intermediaries and hate speech: Fostering digital citizenship for our information age. Boston University Law Review, 91, 1436-1460.
- Chung, Y. L., Kuzmenko, E., Tekiroglu,S., & Guerini, M. (2019). CONAN-COunter NArratives through Nich- esourcing: A multilingual dataset of responses to fight online hate speech. In P. Nakov & A. Palmer (Eds.), Proceedings of the 57th annual meeting of the Association for Computational Linguistics (pp. 2819-2829). Association for Computational Linguis- tics. https://doi.org/10.18653/v1/P19-1271 Council of the European Union. (2008). Council Frame- work Decision 2008/913/JHA of 28 November 2008 on combating certain forms and expressions of racism and xenophobia by means of criminal law. Official Journal of the European Union, 51(L 328/55). https://eur-lex.europa.eu/legal-content/EN/TXT/ ?uri=celex%3A32008F0913 · DOI
- Culpeper, J. (2021). Impoliteness and hate speech: Com- pare and contrast. Journal of Pragmatics, 179, 1-11 https://doi.org/10.1016/j.pragma.2021.04.019 · DOI
- Ernst, J., Schmitt, J. B., Rieger, D., Beier, A. K., Vorderer, P., Bente, G., & Roth, H. J. (2017). Hate beneath the counter speech? A qualitative content analysis of user comments on YouTube related to counter speech videos. Journal for Deradicalization, 10, 1-49.
- Fortuna, P., & Nunes, S. (2018). A survey on automatic detection of hate speech in text. ACM Computing Sur- veys (CSUR), 51(4), 1-30.
- Gelber, K. (2019). Differentiating hate speech: A systemic discrimination approach. Critical Review of Interna- tional Social and Political Philosophy, 24(4), 393-414. https://doi.org/10.1080/13698230.2019.1576006 · DOI
- Maity, S. K., Chakraborty, A., Goyal, P., & Mukherjee, A. (2018). Opinion conflicts: An effective route to detect incivility in twitter. In K. Karahalios, A. Monroy- Hernández, A. Lampinen, & G. Fitzpatrick (Eds.), Proceedings of ACM on Human-Computer Interac- tion (Vol. 2, pp. 1-27). Association for Computing Machinery.
- Matsuda, M. (1989). Public response to racist speech: Considering the victim's story. Michigan Law Review, 87(8), 2320-2381.
- McGowan, M. K. (2009). Oppressive speech. Aus- tralasian Journal of Philosophy, 87(3), 389-407. https://doi.org/10.1080/00048400802370334 · DOI
- Munger, K. (2017). Tweetment effects on the tweeted: Experimentally reducing racist harassment. Political Behavior, 39(3), 629-649.
- Musolff, A. (2015). Dehumanizing metaphors in UK immi- grant debates in press and online media. Journal of Language Aggression and Conflict, 3(1), 41-56.
- Richards, R. D., & Calvert, C. (2000). Counterspeech 2000: A new look at the old remedy for "bad" speech. BYU Law Review, 2(2), 553-586.
- Schieb, C., & Preuss, M. (2016, June 9-13). Governing hate speech by means of counterspeech on Facebook [Conference session]. 66th ICA annual conference, Fukuoka, Japan.
- Silva, L., Mondal, M., Correa, D., Benevenuto, F., & Weber, I. (2016). Analyzing the targets of hate in online social media. Proceedings of the International AAAI Conference on Web and Social Media, 10(1), 687-690.
- Stephan, W., Stephan, C., & Gudykunst, W. (1999). Anxiety in intergroup relations: A comparison of anxiety/uncertainty management theory and inte- grated threat theory. International Journal of Inter- cultural Relations, 23(4), 613-628.
- Strossen, N. (2018). Hate: Why we should resist it with free speech, not censorship. Oxford University Press.
- Weaver, S. (2016). Humor and race. In J. Stone, R. M. Den- nis, P. S. Rizova, A. D. Smith, & X. Hou (Eds.), The Wiley Blackwell encyclopedia of race, ethnicity and nation- alism (pp. 1-3).
- Wohlfeld, M. (2014). Is migration a security issue? In O. Grech & M. Wohlfeld (Eds.), Migration in the Mediterranean: Human rights, security and devel- opment perspectives (pp. 61-77). Mediterranean Academy of Diplomatic Studies.
- Wright, L., Ruths, D., Dillon, K. P., Saleem, H. M., & Benesch, S. (2017). Vectors for counterspeech on Twitter. In Z. Waseem, W. Hui Kyong Chung, D. Hovy, & J. Tetreault (Eds.), Proceedings of the first workshop on abusive language online (pp. 57-62). Association for Computational Linguistics.
- Authors
- Baider, F.
- Affiliations
- University of Cyprus (CY)
- Corresponding author
- Baider, Fabienne
- ORCID
- Baider, Fabienne
- Venue
- Politics and Governance
- Publisher
- Cogitatio
- Published
- 2023-03-20
- Volume
- 11
- Issue
- 2
- ISSN
- 2183-2463 (ISSN-L)
- OpenAlex
- W4327911820
- Cited by
- 33
- References
- 36
- Citations by year
- 2026: 9 · 2025: 15 · 2024: 6 · 2023: 3
- FWCI
- 2.92
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Politics and Governance (publishedVersion, cc-by)
- Social Science Open Access Repository (GESIS – Leibniz Institute for the Social Sciences) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Politics and Governance (submittedVersion, cc-by)
- Funders
- European Commission (2014‐2020, 875263)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Social Sciences › Communication › Social Media and Politics
- Keywords
- online hate speech, counterspeech, counter-narratives, accountability, argumentation strategies, affective strategies, Twitter comments, YouTube comments
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
