📄 Tu chiamale se vuoi emozioni: The impact of emotions on disinformation and sexism identification

Summary

This document is the abstract of an invited keynote at EVALITA 2026 in Bari, not a full research paper. It reports no datasets, experiments or quantitative results. The talk looks at how emotion shapes harmful online content. Its premise is that disinformation usually carries a strong emotional charge. False stories provoke different and more intense feelings than accurate reporting, and readers favour high-arousal items that grab attention and stay in memory. Emotional bias can also distort how people weigh arguments, so they may reject facts they find upsetting.

The core of the talk treats disinformation from a European Union perspective. There, coordinated and often foreign campaigns are seen as information warfare meant to stir negative feelings and polarise democracies. The speaker also examines the reactions triggered by conspiracy narratives, whose language works on readers' subconscious biases. The final part covers sexism identification in memes: which emotions such memes evoke and how physiological measurements can capture those signals. The talk matters because it frames emotion as a central cue for detecting harmful content.

Key points

• The speaker is Paolo Rosso (Universitat Politècnica de València, PRHLT Research Center, ValgrAI). The keynote appears in CEUR Workshop Proceedings for EVALITA 2026 (Feb 26–27, Bari).
• For sexist memes and videos, emotional responses are measured through eye-tracking (fixations, saccades, blinks), heart rate and EEG signals.
• Related projects led by the author are FairTransNLP_Stereotypes, FAKEnHATE-PoC and XAI-DisInfodemics. They are funded by MCIN/AEI and European Union NextGenerationEU/PRTR.
• The author has helped organise over 50 shared tasks at SemEval, CLEF, FIRE, IberEval/IberLEF and EVALITA. The EVALITA tasks include MultiPRIDE, AMI, SardiStance, IronITA and SENTIPOLC.
• The abstract gives no corpus names, models or evaluation figures.