We present an overview of the MM-ArgFallacy2025 shared task on Multimodal Argumentative Fallacy Detection and Classification in Political Debates, co-located with the 12th Workshop on Argument Mining at ACL 2025. The task focuses on identifying and classifying argumentative fallacies across three input modes: text-only, audio-only, and multimodal (text+audio), offering both binary detection (AFD) and multi-class classification (AFC) subtasks. The dataset comprises 18,925 instances for AFD and 3,388 instances for AFC, from the MM-USED-Fallacy corpus on U.S. presidential debates, annotated for six fallacy types: Ad Hominem, Appeal to Authority, Appeal to Emotion, False Cause, Slippery Slope, and Slogan. A total of 5 teams participated: 3 on classification and 2 on detection. Participants employed transformer-based models, particularly RoBERTa variants, with strategies including prompt-guided data augmentation, context integration, specialised loss functions, and various fusion techniques. Audio processing ranged from MFCC features to state-of-the-art speech models. Results demonstrated textual modality dominance, with best text-only performance reaching 0.4856 F1-score for classification and 0.34 for detection. Audio-only approaches underperformed relative to text but showed improvements over previous work, while multimodal fusion showed limited improvements. This task establishes important baselines for multimodal fallacy analysis in political discourse, contributing to computational argumentation and misinformation detection capabilities.

Multimodal Argument Mining (MAM) is a recent area of research aiming to extend argument analysis and improve discourse understanding by incorporating multiple modalities. Initial results confirm the importance of paralinguistic cues in this field. However, the research community still lacks a comprehensive platform where results can be easily reproduced, and methods and models can be stored, compared, and tested against a variety of benchmarks. To address these challenges, we propose MAMKit, an open, publicly available, PyTorch toolkit that consolidates datasets and models, providing a standardized platform for experimentation. MAMKit also includes some new baselines, designed to stimulate research on text and audio encoding and fusion for MAM tasks. Our initial results with MAMKit indicate that advancements in MAM require novel annotation processes to encompass auditory cues effectively.

Recent advances in NLP suggest that some tasks, such as argument detection and relation classification, are better framed in a multimodal perspective. We propose multimodal argument mining for argumentative fallacy classification in political debates. To this end, we release the first corpus for multimodal fallacy classification. Our experiments show that the integration of the audio modality leads to superior classification performance. Our findings confirm that framing fallacy classification as a multimodal task is essential to capture paralinguistic aspects of fallacious arguments.

We propose a study on multimodal argument mining in the domain of political debates. We collate and extend existing corpora and provide an initial empirical study on multimodal architectures, with a special emphasis on input encoding methods. Our results provide interesting indications about future directions in this important domain.