Vol. 6 No. 1 (2026): January−June 2026
Articles

Incremental Contribution of Linguistic, Behavioral, and Propagation Signals to Early Detection of LLM-Generated Coordinated Inauthentic Campaigns

View PDF
  • Minghua Deng,
  • Hao Fu,
  • Yunqiang Sun
Minghua Deng
Computational Data Science, Carnegie Mellon University, PA, USA
Hao Fu
Electronic Science and Technology, Beijing University of Posts and Telecommunications, Beijing, China
Yunqiang Sun
Computer Science and Technology, University of Jinan, Shandong, China

Published 2026-06-29

Keywords

  • coordinated inauthentic behavior,
  • machine-generated text detection,
  • early detection,
  • propagation graphs

How to Cite

Deng, M., Fu, H., & Sun, Y. (2026). Incremental Contribution of Linguistic, Behavioral, and Propagation Signals to Early Detection of LLM-Generated Coordinated Inauthentic Campaigns. Spectrum of Research, 6(1). Retrieved from https://spectrumofresearch.com/index.php/sr/article/view/54

Copyright (c) 2026 Spectrum of Research

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

Large language models lower the cost of producing fluent text at scale, raising concern that coordinated inauthentic campaigns can be seeded and amplified faster than moderators can react. Detection in the first hours after a campaign surface is the operationally relevant regime, yet the relative usefulness of different evidence types in that regime is not well quantified. This study reports an ablation-style comparison of three signal channels — linguistic traces of machine generation, temporal behavioral coordination among accounts, and propagation-graph structure — for early detection of such campaigns. Three real public collections (the User Preference-aware Fake News Detection dataset, a labeled subset of a large graph-structured bot benchmark, and two machine-generated-text benchmarks) supply the signals, and a documented state-actor troll archive supports an attribution case study; a single off-the-shelf classifier is applied to every feature configuration so that differences reflect signal content rather than architecture. Fusing all three channels yields a moderate improvement, reaching F1 0.793 on the larger collection against 0.712 for the strongest single channel. Marginal contributions are uneven and window-dependent: linguistic evidence dominates the earliest window, while propagation structure is weakest early and strengthens as cascades grow. The pattern argues for window-aware weighting rather than uniform fusion.

View PDF

References

  1. [1] Zellers, R., Holtzman, A., Rashkin, H., Bisk, Y., Farhadi, A., Roesner, F., & Choi, Y. (2019). Defending against neural fake news. Advances in Neural Information Processing Systems, 32, 9054–9065.
  2. [2] Pacheco, D., Hui, P.-M., Torres-Lugo, C., Truong, B. T., Flammini, A., & Menczer, F. (2021). Uncovering coordinated networks on social media: Methods and case studies. Proceedings of the International AAAI Conference on Web and Social Media, 15, 455–466.
  3. [3] Mitchell, E., Lee, Y., Khazatsky, A., Manning, C. D., & Finn, C. (2023). DetectGPT: Zero-shot machine-generated text detection using probability curvature. Proceedings of the 40th International Conference on Machine Learning, 202, 24950–24962.
  4. [4] Bian, T., Xiao, X., Xu, T., Zhao, P., Huang, W., Rong, Y., & Huang, J. (2020). Rumor detection on social media with bi-directional graph convolutional networks. Proceedings of the AAAI Conference on Artificial Intelligence, 34(1), 549–556.
  5. [5] Bao, G., Zhao, Y., Teng, Z., Yang, L., & Zhang, Y. (2024). Fast-DetectGPT: Efficient zero-shot detection of machine-generated text via conditional probability curvature. Proceedings of the 12th International Conference on Learning Representations (ICLR).
  6. [6] Uchendu, A., Le, T., Shu, K., & Lee, D. (2020). Authorship attribution for neural text generation. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), 8384–8395.
  7. [7] Kumarage, T., Garland, J., Bhattacharjee, A., Trapeznikov, K., Ruston, S., & Liu, H. (2023). Stylometric detection of AI-generated text in Twitter timelines. arXiv. https://arxiv.org/abs/2303.03697
  8. [8] Nizzoli, L., Tardelli, S., Avvenuti, M., Cresci, S., & Tesconi, M. (2021). Coordinated behavior on social media in 2019 UK general election. Proceedings of the International AAAI Conference on Web and Social Media, 15, 443–454.
  9. [9] Feng, S., Wan, H., Wang, N., Li, J., & Luo, M. (2021). TwiBot-20: A comprehensive Twitter bot detection benchmark. Proceedings of the 30th ACM International Conference on Information & Knowledge Management (CIKM), 4485–4494.
  10. [10] Lu, Y.-J., & Li, C.-T. (2020). GCAN: Graph-aware co-attention networks for explainable fake news detection on social media. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), 505–514.
  11. [11] Nguyen, V.-H., Sugiyama, K., Nakov, P., & Kan, M.-Y. (2020). FANG: Leveraging social context for fake news detection using graph representation. Proceedings of the 29th ACM International Conference on Information & Knowledge Management (CIKM), 1165–1174.
  12. [12] Shu, K., Cui, L., Wang, S., Lee, D., & Liu, H. (2019). dEFEND: Explainable fake news detection. Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD), 395–405.
  13. [13] Dou, Y., Shu, K., Xia, C., Yu, P. S., & Sun, L. (2021). User preference-aware fake news detection. Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), 2051–2055.
  14. [14] Shu, K., Mahudeswaran, D., Wang, S., Lee, D., & Liu, H. (2020). 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.
  15. [15] Feng, S., Tan, Z., Wan, H., Wang, N., Chen, Z., Zhang, B., Zheng, Q., … Luo, M. (2022). TwiBot-22: Towards graph-based Twitter bot detection. Advances in Neural Information Processing Systems, 35 (Datasets and Benchmarks Track).
  16. [16] Dugan, L., Hwang, A., Trhlík, F., Zhu, A., Ludan, J. M., Xu, H., Ippolito, D., & Callison-Burch, C. (2024). RAID: A shared benchmark for robust evaluation of machine-generated text detectors. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), 12463–12492.
  17. [17] Wang, Y., Mansurov, J., Ivanov, P., Su, J., Shelmanov, A., Tsvigun, A., Whitehouse, C., Afzal, O. M., Mahmoud, T., Sasaki, T., Arnold, T., Aji, A. F., Habash, N., Gurevych, I., & Nakov, P. (2024). M4: Multi-generator, multi-domain, and multi-lingual black-box machine-generated text detection. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL), 1369–1407.
  18. [18] Uchendu, A., Ma, Z., Le, T., Zhang, R., & Lee, D. (2021). TURINGBENCH: A benchmark environment for Turing test in the age of neural text generation. Findings of the Association for Computational Linguistics: EMNLP 2021, 2001–2016.
  19. [19] Ma, J., Gao, W., Mitra, P., Kwon, S., Jansen, B. J., Wong, K.-F., & Cha, M. (2016). Detecting rumors from microblogs with recurrent neural networks. Proceedings of the 25th International Joint Conference on Artificial Intelligence (IJCAI), 3818–3824.
  20. [20] Ma, J., Gao, W., & Wong, K.-F. (2017). Detect rumors in microblog posts using propagation structure via kernel learning. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL), 708–717.
  21. [21] Li, Y., Li, Q., Cui, L., Bi, W., Wang, Z., Wang, L., Yang, L., Shi, S., & Zhang, Y. (2024). MAGE: Machine-generated text detection in the wild. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), 36–53.
  22. [22] Guo, B., Zhang, X., Wang, Z., Jiang, M., Nie, J., Ding, Y., Yue, J., & Wu, Y. (2023). How close is ChatGPT to human experts? Comparison corpus, evaluation, and detection. arXiv. https://arxiv.org/abs/2301.07597
  23. [23] Monti, F., Frasca, F., Eynard, D., Mannion, D., & Bronstein, M. M. (2019). Fake news detection on social media using geometric deep learning. arXiv. https://arxiv.org/abs/1902.06673
  24. [24] Fagni, T., Falchi, F., Gambini, M., Martella, A., & Tesconi, M. (2021). TweepFake: About detecting deepfake tweets. PLOS ONE, 16(5), e0251415