Elamur Research⌕
☰
home / papers / 2026 / MALicious INTent Dataset and Inoculating LLMs for Enhanced Disinformation Detection

MALicious INTent Dataset and Inoculating LLMs for Enhanced Disinformation Detection

Paper · 2026 · Arkadiusz Modzelewski, Witold Sosnowski, Eleni Papadopulos et al.

Summary

Most English disinformation resources only mark whether content is false. They ignore the motive behind it, even though intent is central to how disinformation is defined. The authors fill this gap with MALINT, an English corpus of online news articles labelled for credibility and for the malicious goals of disinformation authors. Annotators were trained and supervised by fact-checkers from IFCN-accredited organisations. Disinformative articles received multilabel tags from five intent categories.

With this corpus, the authors evaluate 12 language models, ranging from fine-tuned SLMs like BERT to LLMs like Llama 3.3, on identifying intent in both binary and multilabel settings. Drawing on inoculation theory from psychology, the authors also propose intent-based inoculation (IBI). In this prompting strategy, an LLM first reasons about possible malicious intent and only then judges whether a text is disinformation.

Across six datasets, five LLMs and seven languages, IBI consistently improves zero-shot detection. This suggests that modelling intent is a practical way to strengthen disinformation detectors, including for lower-resource languages.

Key points

  • MALINT corpus: articles come from roughly 50 sources that experts grouped as Reliable, Unreliable or Mixed/Biased, with source labels hidden from annotators. Each article was labelled Credible Information, Disinformation or Hard-to-say, and the last group was left out of experiments.
  • Intent taxonomy: the five multilabel categories are UCPI, CPV, UIOA, PSSA and PASV. They cover attacks on public institutions, political persuasion, attacks on international alliances, stoking social divisions, and anti-science messaging.
  • Intent classification: fine-tuned SLMs perform best in the multilabel setting, reaching a weighted F1 of 82.1%. In the binary per-category setup, LLMs beat the SLMs on three of the five intents.
  • IBI results: IBI raises zero-shot disinformation detection by about 9% on average over five English datasets and the tested LLMs. Gains are even larger in other languages, including Estonian and Polish. The evaluation also used genre splits (articles vs. posts) and temporal splits (before vs. after the models' knowledge cutoff).
  • Release: the dataset, including annotations from every annotation step, is published under CC BY 4.0. Prompts and code are available at github.com/ArkadiusDS/MALINT.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of PaduaPadua · Italy18 works · 4,260 citations🎓Politecnico di TorinoTurin · Italy11 works · 18 citations🎓Polish-Japanese Academy of Information TechnologyWarsaw · Poland2 works · 74 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksSocial Sciences · 1,954 worksInformation Systems · 1,730 worksSociology and Political Science · 626 works

Cite

Cite as Arkadiusz Modzelewski, Witold Sosnowski, Eleni Papadopulos, Elisa Sartori, Tiziano Labruna, Giovanni Da San Martino and Adam Wierzbicki. 2026. MALicious INTent Dataset and Inoculating LLMs for Enhanced Disinformation Detection. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.eacl-long.144

@inproceedings{modzelewski2026malicious,
  title = {{MALicious INTent Dataset and Inoculating LLMs for Enhanced Disinformation Detection}},
  author = {Arkadiusz Modzelewski, Witold Sosnowski, Eleni Papadopulos, Elisa Sartori, Tiziano Labruna, Giovanni {Da San Martino} and Adam Wierzbicki},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.144},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.144}
}

References

53 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Hadeer Ahmed, Issa Traore, and Sherif Saad. 2017. Detection of online fake news using n-gram analysis and machine learning techniques. In Intelligent, Secure, and Dependable Systems in Distributed and Cloud Environments: First International Conference, IS-DDC 2017, Vancouver, BC, Canada, October 26-… find
  2. Hadeer Ahmed, Issa Traore, and Sherif Saad. 2018. Detecting opinion spams and fake news using text classification. Security and Privacy, 1(1):e9. find
  3. Naomi Appelman, Stephan Dreyer, Pranav Manjesh Bidare, and Keno C Potthast. 2022. Truth, intention and harm: Conceptual challenges for disinformation-targeted governance. Internet Policy Review. find
  4. Yejin Bang, Samuel Cahyawijaya, Nayeon Lee, Wenliang Dai, Dan Su, Bryan Wilie, Holy Lovenia, Ziwei Ji, Tiezheng Yu, Willy Chung, et al. 2023. A multi-task, multilingual, multimodal evaluation of chatgpt on reasoning, hallucination, and interactivity. In Proceedings of the 13th International Joint Co… find
  5. Rachit Bansal, William Scott Paka, Nidhi, Shubhashis Sengupta, and Tanmoy Chakraborty. 2021. Combining exogenous and endogenous signals with a semi-supervised co-attention network for early detection of covid-19 fake tweets. In Pacific-Asia conference on knowledge discovery and data mining, pages 18… find
  6. John Blitzer, Ryan McDonald, and Fernando Pereira. 2006. Domain adaptation with structural correspondence learning. In Proceedings of the 2006 conference on empirical methods in natural language processing, pages 120–128.
  7. Dallas Card, Peter Henderson, Urvashi Khandelwal, Robin Jia, Kyle Mahowald, and Dan Jurafsky. 2020. With little power comes great responsibility. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 9263–9274. find
  8. Mark Conner and Paul Norman. 2022. Understanding the intention-behavior gap: The role of intention strength. Frontiers in Psychology, 13. find
  9. Limeng Cui and Dongwon Lee. 2020. Coaid: Covid-19 healthcare misinformation dataset. arXiv preprint arXiv:2006.00885. find
  10. Madeleine de Cock Buning. 2018. A multi-dimensional approach to disinformation: Report of the independent high level group on fake news and online disinformation. Publications Office of the European Union. find
  11. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. BERT: pre-training of deep bidirectional transformers for language understanding. CoRR, abs/1810.04805.
  12. Thomas G Dietterich. 1998. Approximate statistical tests for comparing supervised classification learning algorithms. Neural computation, 10(7):1895–1923. find
  13. Rotem Dror, Gili Baumer, Segev Shlomov, and Roi Reichart. 2018. The hitchhiker’s guide to testing statistical significance in natural language processing. In Proceedings of the 56th annual meeting of the association for computational linguistics (volume 1: Long papers), pages 1383–1392. find
  14. European Commission. 2022a. Communication from the commission to the european parliament, the council, the european economic and social committee and the committee of the regions on the european democracy action plan. Publications Office of the European Union. Available at: https://eur-lex.europa.eu… ↗
  15. European Commission. 2022b. The strengthened code of practice on disinformation 2022. Publications Office of the European Union. Available at: https://digital-strategy.ec.europa.eu/en/library/2022-strengthened-code-practice-disinformation. ↗
  16. Ankur Gupta, Yash Varun, Prarthana Das, Nithya Muttineni, Parth Srivastava, Hamim Zafar, Tanmoy Chakraborty, and Swaprava Nath. 2021. Truthbot: An automated conversational tool for intent learning, curated information presenting, and fake news alerting. Preprint, arXiv:2102.00509. — [arXiv:2102.00509](https://arxiv.org/abs/2102.00509) find
  17. Michael Hameleers. 2022. Disinformation as a context-bound phenomenon: toward a conceptual clarification integrating actors, intentions and techniques of creation and dissemination. Communication Theory, 33(1):1–10. find
  18. Michael Hameleers. 2023. Disinformation as a context-bound phenomenon: Toward a conceptual clarification integrating actors, intentions and techniques of creation and dissemination. Communication Theory, 33(1):1–10. find
  19. Pengcheng He, Jianfeng Gao, and Weizhu Chen. 2021a. Debertav3: Improving deberta using electra-style pre-training with gradient-disentangled embedding sharing. Preprint, arXiv:2111.09543.
  20. Pengcheng He, Xiaodong Liu, Jianfeng Gao, and Weizhu Chen. 2021b. Deberta: Decoding-enhanced bert with disentangled attention. In International Conference on Learning Representations.
  21. Junaed Younus Khan, Md Tawkat Islam Khondaker, Sadia Afroz, Gias Uddin, and Anindya Iqbal. 2021. A benchmark study of machine learning models for online fake news detection. Machine Learning with Applications, 4:100032. find
  22. Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35:22199– 22213.
  23. Julia Kruk, Jonah Lubin, Karan Sikka, Xiao Lin, Dan Jurafsky, and Ajay Divakaran. 2019. Integrating text and image: Determining multimodal document intent in Instagram posts. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Con… find
  24. João A Leite, Olesya Razuvayevskaya, Kalina Bontcheva, and Carolina Scarton. 2024. Euvsdisinfo: a dataset for multilingual detection of pro-kremlin disinformation in news articles. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, pages 5380–5384. find
  25. Stephan Lewandowsky, Ullrich KH Ecker, and John Cook. 2017. Beyond misinformation: Understanding and coping with the “post-truth” era. Journal of applied research in memory and cognition, 6(4):353– 369. find
  26. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized BERT pretraining approach. CoRR, abs/1907.11692.
  27. Jason Lucas, Adaku Uchendu, Michiharu Yamashita, Jooyoung Lee, Shaurya Rohatgi, and Dongwon Lee. 2023. Fighting fire with fire: The dual role of llms in crafting and detecting elusive disinformation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 142… find
  28. Adyasha Maharana, Quan Tran, Franck Dernoncourt, Seunghyun Yoon, Trung Bui, Walter Chang, and Mohit Bansal. 2022. Multimodal intent discovery from livestream videos. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 476–489, Seattle, United States. Association for Compu… find
  29. William J McGuire. 1964. Inducing resistance to persuasion. some contemporary approaches. CC Haaland and WO Kaelber (Eds.), Self and Society. An Anthology of Readings, Lexington, Mass.(Ginn Custom Publishing) 1981, pp. 192-230. 2023. Multilingual multifaceted understanding of online news in terms of… find
  30. Arkadiusz Modzelewski, Giovanni Da San Martino, Pavel Savov, Magdalena Wilczyńska, and Adam Wierzbicki. 2024. Mipd: Exploring manipulation and intention in a novel corpus of polish disinformation. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 19769… find
  31. Jon Roozenbeek, Sander van der Linden, and Thomas Nygren. 2020. Prebunking interventions based on “inoculation” theory can reduce susceptibility to misinformation across cultures. The Harvard Kennedy School (HKS) Misinformation Review. find
  32. Arkadiusz Modzelewski, Witold Sosnowski, Tiziano Labruna, Adam Wierzbicki, and Giovanni Da San Martino. 2025. Pcot: Persuasion-augmented chain of thought for detecting fake news and social media disinformation. arXiv preprint arXiv:2506.06842. find
  33. Kevin P Murphy. 2018. Machine learning: A probabilistic perspective (adaptive computation and machine learning series). The MIT Press: London, UK. find
  34. Preslav Nakov, Jisun An, Haewoon Kwak, Muhammad Arslan Manzoor, Zain Muhammad Mujahid, and Husrev T Sencar. 2024. A survey on predicting the factuality and the bias of news media. In ACL (Findings). find
  35. Hai-Long Nguyen, Thi-Kieu-Trang Pham, Thai-Son Le, Tan-Minh Nguyen, Thi-Hai-Yen Vuong, and Ha-Thanh Nguyen. 2023. Rmdm: A multilabel fakenews dataset for vietnamese evidence verification. arXiv preprint arXiv:2309.09071. — [arXiv:2309.09071](https://arxiv.org/abs/2309.09071) find
  36. James Pamment and Anneli Lindvall Kimber. 2021. Fact-checking and debunking: a best practice guide to dealing with disinformation. NATO Strategic Communication Centre of Excellence. find
  37. Tsung-Hsuan Pan, Chung-Chi Chen, Hen-Hsen Huang, and Hsin-Hsi Chen. 2024. Enhancing society-undermining disinformation detection through fine-grained sentiment analysis pre-finetuning. In Findings of the Association for Computational Linguistics: EACL 2024, pages 1371–1377. find
  38. Kellin Pelrine, Anne Imouza, Camille Thibault, Meilina Reksoprodjo, Caleb Gupta, Joel Christoph, Jean-François Godbout, and Reihaneh Rabbany. 2023. Towards reliable misinformation mitigation: Generalization, uncertainty, and gpt-4. In Proceedings of the 2023 Conference on Empirical Methods in Natura… find
  39. Michael Pfau, Bobi Ivanov, Brian Houston, Michel Haigh, Jeanetta Sims, Eileen Gilchrist, Jason Russell, Shelley Wigley, Jackie Eckstein, and Natalie Richert. 2005. Inoculation and mental processing: The instrumental role of associative networks in the process of resistance to counterattitudinal infl… find
  40. Jakub Piskorski, Nicolas Stefanovitch, Nikolaos Nikolaidis, Giovanni Da San Martino, and Preslav Nakov. find
  41. Victor Sanh, Lysandre Debut, Julien Chaumond, and Thomas Wolf. 2019. Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. ArXiv, abs/1910.01108.
  42. MG Sessa. 2023. Connecting the disinformation dots: Insights, lessons, and guidance from 20 eu member states. EU Disinfolab, Friedrich Naumann Foundation for Freedom. find
  43. Kai Shu, Amrita Bhattacharjee, Faisal Alatawi, Tahora H Nazer, Kaize Ding, Mansooreh Karami, and Huan Liu. 2020a. Combating disinformation in a social media age. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(6):e1385. find
  44. Kai Shu, Deepak Mahudeswaran, Suhang Wang, Dongwon Lee, and Huan Liu. 2020b. Fakenewsnet: A data repository with news content, social context, and spatiotemporal information for studying fake news on social media. Big data, 8(3):171–188. find
  45. Kai Shu, Amy Sliva, Suhang Wang, Jiliang Tang, and Huan Liu. 2017. Fake news detection on social media: A data mining perspective. ACM SIGKDD explorations newsletter, 19(1):22–36. find
  46. Witold Sosnowski, Arkadiusz Modzelewski, Kinga Skorupska, Jahna Otterbacher, and Adam Wierzbicki. 2024. Eu disinfotest: a benchmark for evaluating language models’ ability to detect disinformation narratives. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 14702– 1472… find
  47. Cecilie S Traberg, Jon Roozenbeek, and Sander van der Linden. 2022. Psychological inoculation against misinformation: Current evidence and future directions. The ANNALS of the American Academy of Political and Social Science, 700(1):136–151. find
  48. Antje von Suchodoletz and Anja Achtziger. 2011. Intentions and their limits. Social Psychology, 42:85–92. find
  49. Bing Wang, Ximing Li, Changchun Li, Bo Fu, Songwen Pei, and Shengsheng Wang. 2024a. Why misinformation is created? detecting them by integrating intent features. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, pages 2304–2314. find
  50. Bing Wang, Ximing Li, Changchun Li, Bo Fu, Songwen Pei, and Shengsheng Wang. 2024b. Why misinformation is created? detecting them by integrating intent features. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, CIKM ’24, page 2304–2314, New York, NY, U… find
  51. William Yang Wang. 2017. “liar, liar pants on fire”: A new benchmark dataset for fake news detection. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 422–426.
  52. Hua Xu, Hanlei Zhang, and Ting-En Lin. 2023. Intent Recognition, pages 7–29. Springer Nature Singapore, Singapore. find
  53. Xinyi Zhou, Kai Shu, Vir V. Phoha, Huan Liu, and Reza Zafarani. 2022. “this is fake! shared it by mistake”:assessing the intent of fake news spreaders. In Proceedings of the ACM Web Conference 2022, WWW ’22, page 3685–3694. ACM. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/modzelewski-2026-malicious-intent-dataset-inoculating-llms/