"The idea that #disinformation and #hate can only be combated through direct #counterspeech on these platforms assumes that genuine, fair discourse takes place there. But that is precisely not the case. The #algorithms of these platforms are optimized to maximize #engagement/attention - and this […]
Example of the functioning of our Speech Concept Bottleneck Models #SCBM
📄https://doi.org/10.1016/j.ipm.2025.104309
#XAI #explainability #interpretability #ML #AI #hatespeech #hate #counterspeech #toxic
✨ What’s inside?
🔹 Distilling knowledge from #LLMs into transparent #SCBM
🔹 Using domain-specific human-understandable adjectives as concepts
🔹 A novel class-discriminative loss for #interpretability
🔹 Applied to #hate & #counterspeech recognition
In our new paper "Distilling knowledge from large language models: A #concept #bottleneck #model for #hate and #counterspeech recognition" we introduce Speech Concept Bottleneck Models #SCBM - a step toward #interpretable #LLM.
📄https://doi.org/10.1016/j.ipm.2025.104309
Taxonomy of Strategies to Counter Toxic Online Content
New research groups 25 response tactics for toxic online content into five evidence‑based categories, offering a clear framework for moderators and AI developers. Read more: getnews.me/taxonomy-of-strategies-t... #counterspeech #onlinediscourse #toxiconline
A researcher in our project on counterspeech wrote this great critique of the concept from an intersectional feminist perspective.
https://doi.org/10.1080/14680777.2025.2528068
#counterspeech #moderation #meta #hatespeech
🗣️ Discover "NLP for Counterspeech against Hate and Misinformation" at #ACL2025NLP! This tutorial bridges CS, social sciences & policy, exploring among others the role of LLMs.
#Counterspeech #HateSpeech #Misinformation #NLProc #NLG #ResponsibleAI
2025.aclweb.org/program/tuto...
📢 In a new online field experiment, we find that #counterspeech with perspective-based messages can reduce online hate speech, and its 🌟amplification🌟
with a large research team across #ETHZurich @ipz.bsky.social @uclspp.bsky.social and beyond
www.nature.com/articles/s41...
Die Schattenseiten der Anonymität in sozialen Medien - Counterspeech reloaded - Trolle, Tweets & Tabubrüche wix.to/o3THpXA #Counterspeech #HassImNetz
Counterspeech's impact varies; real-world, scalable tests are needed. Interdisciplinary collaboration is key to automate effective hate mitigation. #counterspeech
The Roots of Counterspeech: A Review of Social and Technical Perspectives #Technology #SocialandEthicalImplications #Counterspeech #SocialMedia #TechEthics
Counterspeech counters online hate via supportive rebuttals. This review bridges social and computer science to enhance automated generation. #counterspeech
📄 Based on:
ELF22: A Context-based Counter Trolling Dataset to Combat Internet Trolls
Lee et al., 2022
🔗 arxiv.org/pdf/2208.0...
#TrollTactics5 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 5 – Final Recap: How to Fight Trolls with Data, Strategy, and AI
#TrollTactics5 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
We won’t moderate our way out of trolling.
But we can design systems that fight back intelligently.
AI trained on tactics—not outrage—might be the first step.
#TrollTactics4 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 4 – How AI Learns to Clap Back
#TrollTactics4 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Know what kind of troll you’re dealing with—then choose the right counter.
Trolling isn’t just about tone.
It’s about control.
#TrollTactics3 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 3 – Overt vs. Covert Trolls: Why It Matters
#TrollTactics3 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
The goal isn’t to destroy the troll.
It’s to protect the space.
These 7 tactics give you options—so you can respond, or not, on your terms.
#TrollTactics2 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 2 – The 7 Ways to Shut a Troll Down
#TrollTactics2 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
“Don’t feed the trolls” is a useful default.
But it’s not a strategy.
Knowing how to respond —and when not to—is how communities stay resilient.
#TrollTactics1 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 1 – Why “Don’t Feed the Trolls” Doesn’t Cut It
#TrollTactics1 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Based on:
📄 ELF22: A Context-based Counter Trolling Dataset to Combat Internet Trolls
by Huije Lee et al. (2022)
🔗 arxiv.org/pdf/2208.0...
#TrollTactics0 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
Thread 0 – Trolls, Tactics, and Counter Speech
Today is Sunday, so I thought, why not keeping it 'light' :)
We'll do trolls and tactics 🧵
#TrollTactics0 #OnlineHarassment #CounterSpeech #DigitalResilience #Trolling
(Deutscher) Journalismus fühlt sich grad auch bissi an wie #Counterspeech gegen Selbstermächtigung von Leuten, die „vorher nix zu sagen“ hatten…
Das ist ein guter Ansatzpunkt für Gespräche mit noch-nicht-ganz-Verlorenen: nachfragen, was genau sie an ihm gut finden. Wie die Welt aussähe, wenn es nur Leute wie Musk gäbe. Wer davon profitiert und ob man wirklich meint, am Ende zu den Profiteuren zu gehören.
#CounterSpeech #Rhetorik
Auch wenn's schwerfällt, wenn man gerade der Debatte im BT folgt:
Bitte NICHT die entgrenzten Provokationen & Falschbehauptungen wiederholen!!
Es setzt sich nämlich auch wenn man empört ist, fest. Bitte kontert dagegen OHNE WIEDERHOLUNG mit Fakten - und Menschlichkeit.
#Rhetorik #CounterSpeech
Guter Artikel, der im Wesentlichen wiedergibt, was ich in meinen Workshops u.a. auch vermittle: Fragen stellen, gemeinsame Ziele statt Positionen, über Emotionen gehen usw.
#CounterSpeech #Argumentation #Rhetorik
New blog 🎉🖊️
Read a wrap-up of the multidisciplinary #HateSpeech workshop that took place on 14Nov
at Trinity College, Cambridge. I delivered a talk on mitigation strategies like quarantining & #Counterspeech. Now up on the
@mctd.bsky.social website:
bit.ly/4f92uFQ
@crasshlive.bsky.social
