Overview of EXIST 2026: Physiological Data for Multimodal Sexism Characterization in Social Media

Abstract

This paper presents the EXIST 2026 Lab on sexism detection and categorization in social media, held at the CLEF 2026 conference, and marks the sixth edition of the EXIST Shared Task. EXIST 2026 further expands the study of online sexism by focusing on complex multimedia formats where sexist meaning may emerge from the interaction between visual, textual, temporal, and contextual cues. The lab comprises six subtasks in two languages, English and Spanish, organized around three core objectives: sexism identification, source intention detection, and sexism categorization. These objectives are applied across two multimedia formats: images, in the form of memes, and videos, in the form of TikToks. In contrast to previous editions, EXIST 2026 focuses exclusively on multimedia content and introduces a Human-Centered AI perspective by enriching the data with sensor-based information, including eye-tracking, heart-rate, and EEG signals collected from subjects exposed to potentially sexist content. As in previous editions, EXIST 2026 adopts the “Learning With Disagreement” paradigm, using annotations from multiple annotators to represent diverse and sometimes conflicting perceptions of sexism. This edition extends that paradigm by exploring how conscious labels and unconscious physiological or behavioral responses can jointly contribute to the development and evaluation of sexism detection systems. This overview describes the task design, datasets, evaluation methodology, participating systems, and results of EXIST 2026. The lab registered 247 teams from 26 countries, with 122 teams from 16 countries submitting runs, and a total of 1,303 runs processed. Warning: Some of the examples included in this paper may contain offensive language and explicit descriptions of sexist behavior, which may be disturbing to the reader.