Text classification methods have been widely investigated as a way to detect content of low credibility: fake news, social media bots, propaganda, etc. Quite accurate models (likely based on deep neural networks) help in moderating public electronic platforms and often cause content creators to face rejection of their submissions or removal of already published texts. Having the incentive to evade further detection, content creators try to come up with a slightly modified version of the text (known as an attack with an adversarial example) that exploit the weaknesses of classifiers and result in a different output. Here we systematically test the robustness of common text classifiers against available attacking techniques and discover that, indeed, meaning-preserving changes in input text can mislead the models. The approaches we test focus on finding vulnerable spans in text and replacing individual characters or words, taking into account the similarity between the original and replacement content. We also introduce BODEGA: a benchmark for testing both victim models and attack methods on four misinformation detection tasks in an evaluation framework designed to simulate real use-cases of content moderation. The attacked tasks include (1) fact checking and detection of (2) hyperpartisan news, (3) propaganda and (4) rumours. Our experimental results show that modern large language models are often more vulnerable to attacks than previous, smaller solutions, e.g. attacks on GEMMA being up to 27\% more successful than those on BERT. Finally, we manually analyse a subset adversarial examples and check what kinds of modifications are used in successful attacks.
The proliferation of fake news, i.e., news intentionally spread for misinformation, poses a threat to individuals and society. Despite various fact-checking websites such as PolitiFact, robust detection techniques are required to deal with the increase in fake news. Several deep learning models show promising results for fake news classification, however, their black-box nature makes it difficult to explain their classification decisions and quality-assure the models. We here address this problem by proposing a novel interpretable fake news detection framework based on the recently introduced …
LLMs as annotators of credibility assessment in Danish asylum decisions: evaluating classification performance and errors beyond aggregated metrics
Visual Analysis and Perception Lab（视觉分析与感知实验室）；Pioneer Center for AI（先锋人工智能中心）；Center of Excellence for Global Mobility Law（全球移动法律卓越中心）；Department of Computer Science（计算机科学系）
Neural Network Architecture for Credibility Assessment of Textual Claims
Language Technologies Research Centre (LTRC)（语言技术研究中心（LTRC））；Kohli Center on Intelligent Systems (KCIS)（科利智能系统中心（KCIS））；International Institute of Information Technology, Hyderabad（海得拉巴国际信息技术学院）
German Research Center for Artificial Intelligence（德国人工智能研究中心）；pc GmbH Neue Kommunikation（3pc新通信有限公司）
The increasingly rapid spread of information about COVID-19 on the web calls for automatic measures of quality assurance. In that context, we check the credibility of news content using selected linguistic features. We present two empirical studies to evaluate the usability of graphical interfaces that offer such credibility assessment. In a moderated qualitative interview with six participants, we identify rating scale, sub-criteria and algorithm authorship as important predictors of the usability. A subsequent quantitative online survey with 50 participants reveals a conflict between transpa…
This project demonstrated a methodology to estimating cooperate credibility with a Natural Language Processing approach. As cooperate transparency impacts both the credibility and possible future earnings of the firm, it is an important factor to be considered by banks and investors on risk assessments of listed firms. This approach of estimating cooperate credibility can bypass human bias and inconsistency in the risk assessment, the use of large quantitative data and neural network models provides more accurate estimation in a more efficient manner compare to manual assessment. At the beginn…
CrediBench: Building Web-Scale Network Datasets for Information Integrity
McGill University（麦吉尔大学）；Mila - Quebec AI Institute（魁北克人工智能研究所）；University of Oxford（牛津大学）；University of California, Berkeley（加州大学伯克利分校）；AITHYRA Research Institute（AITHYRA研究院）；Université de Montréal（蒙特利尔大学）
A Cross-Cultural Assessment of Human Ability to Detect LLM-Generated Fake News about South Africa
IU International University of Applied Sciences, Germany（国际应用科学大学）；Karlsruhe University of Applied Sciences, Germany（卡尔斯鲁厄应用科学大学）；Applied Information Systems, University of Johannesburg, South Africa（应用信息系统，约翰内斯堡大学）
HarmMetric Eval: Benchmarking Metrics and Judges for LLM Harmfulness Assessment
The State Key Laboratory of Blockchain and Data Security（区块链与数据安全国家重点实验室）；Zhejiang University（浙江大学）；Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security（杭州高新技术区（滨江）区块链与数据安全研究院）；Pretraining Team, xAI（预训练团队，xAI）；Hong Kong University of Science and Technology (GuangZhou)（香港科技大学（广州））
