Tag
An EMNLP Findings 2026 paper introduces a multi-agent framework (story, image, and critic agents) that generates over 9,000 multimodal fake news posts and benchmarks 16 open- and closed-source MLLMs, finding they fall short of human-level detection accuracy, especially at judging image authenticity.
A game similar to Wordle designed to enhance media literacy by focusing on fake news and memes.
The study examines how large language models generate and detect fake news under different scenarios, revealing variations in performance and that refined prompts do not always improve detection.
A US-linked network of fake websites is targeting AI chatbots with Alberta separatism content, aiming to influence AI systems.
BharatGather is a culturally-informed benchmark dataset designed for binary misinformation detection in Indian mass gatherings, comprising 14,646 records to address socio-cultural nuances in automated fake news detection.
A study with 504 participants found that heightened suspicion did not improve detection of AI-generated text, and fake-news accuracy decreased by 10.2 points under sustained exposure, indicating cognitive fatigue in media evaluation.
A study finds that while AI chatbots initially improve fake news detection, prolonged use leads to worse independent performance, highlighting an AI dependency paradox. The research indicates that Socratic questioning AI styles may better engage users in learning to discern truth.
This paper addresses the growing threat of pure-synthesis fake news videos generated by text-to-video models, introducing a new ternary classification task and the first pure-synthesis fake news video dataset (PS-FNVD), along with a Reasoning-guided framework (R-T2V) that achieves state-of-the-art detection accuracy.
This paper presents a multimodal NLP framework that fuses XLM-RoBERTa and CLIP with geospatial and sarcasm features to detect fake news and predict violence-driven mob activity, achieving 98% test accuracy on a 138,256-sample Bangla/English dataset.
An article discussing the ironic situation where AI-generated fake news complains about the threat of AI fake news to real journalism.
Researchers from Kennesaw State University investigate cross-prompt generalization in detecting AI-generated fake news using interpretable linguistic features (lexical diversity, readability, emotion). A random forest classifier trained on one prompting strategy and tested on another achieves AUC values of 0.988–1.000, suggesting these features capture stable, generalizable properties of AI-generated text.
Proposes the CORE framework that endows multimodal large language models with explicit conflict-capturing capability for generalizable manipulation detection, adapting to unseen manipulation types with few or zero samples.
This paper introduces BOUTEF, a large-scale multilingual corpus for studying fake news in Algeria and Tunisia, covering Arabic dialects, Arabizi, French, English, and code-switching. It includes empirical analysis of linguistic strategies and engagement dynamics.
A correction issued by the Department of War clarifies that SpaceX remains a strong and valued partner, refuting false claims in a news article.
A rant about the proliferation of fake references and AI-hallucinated data in tech articles, using examples of a false story about Swedish crows and an inflated claim about code review defect detection.
Donald Trump stated that Israel did not urge him to attack Iran; his decision was driven by the Oct 7, 2023 events and his belief that Iran must not obtain nuclear weapons, while also criticizing media and polls as largely fabricated.