Elamur Research⌕
☰
home / papers / 2026 / DAMASHA: Detecting AI in Mixed Adversarial Texts via Segmentation with Human-interpretable Attribution

DAMASHA: Detecting AI in Mixed Adversarial Texts via Segmentation with Human-interpretable Attribution

Paper · 2026 · L. D. M. S. Sai Teja, N. Siva Gopala Krishna, Ufaq Khan et al.

Summary

The paper addresses mixed-authorship segmentation: locating the points in a text where writing switches between a human and an AI model. It focuses on texts that have been deliberately perturbed to evade detectors. The authors observe that earlier work handled either boundary detection on clean mixed texts or adversarial robustness for whole-document classification, but never both together.

They release MAS, a benchmark that merges the TriBERT and M4GT mixed-text corpora with newly generated material. Every text in the combined corpus is then copied under several character- and punctuation-level attacks.

Their model, Info-Mask, adds a soft attribution mask that reweights token representations using stylometric and perplexity-based cues. The mask also provides Human-Interpretable Attribution (HIA) overlays that let reviewers see why a span was flagged.

The best variant, RMC*, reaches modest strict span-level scores. Most of its errors, however, miss the boundary by only a few tokens. In a human study, the overlays made annotators' decisions faster. The method currently supports only English.

Key points

  • MAS size and sources:
    • 95,859 base texts, multiplied by 7 versions (original plus attacked copies), for 671,013 rows in total.
    • Of the base texts, 46,830 are newly generated with GPT-4o, GPT-4.1, GPT-4.1-mini and DeepSeek-V3.
    • Domains are Reddit, News, Wikipedia, ArXiv and Q&A.
    • Released on Hugging Face as saiteja33/DAMASHA.
  • Attacks and data formats:
    • Listed attacks are misspelling, character substitution, invisible characters, punctuation substitution and upper/lower case swaps.
    • Texts come in Human→AI, AI→Human and sentence-level "Mix" formats, with boundaries annotated at the word level.
  • New metrics: Segment-wise Boundary Detection Accuracy (SBDA) and Segment Precision. Both are computed with IoU span matching at thresholds τ ∈ {0.3, 0.5, 0.7, 0.9}.
  • Main results for RMC with all features:*
    • Strict scores are SBDA 0.45 and SegPre 0.41.
    • Relaxed span accuracy at IoU ≥ 0.5 rises to 0.82.
    • 46% of boundary errors are off by just one token.
  • Interpretability checks:
    • Faithfulness test: masking the top-10% mask-weighted tokens drops SBDA@0.3 from 0.47 to 0.33.
    • Human study: covered 500 samples rated by 2 annotators (κ = 0.68).
    • Effect of overlays: decision time fell from 21.4 s to 8.7 s, and usefulness was rated 4.3 out of 5.

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

🎓Mohamed bin Zayed University of Artificial IntelligenceAbu Dhabi · United Arab Emirates124 works · 5,676 citations🎓National Institute Of Technology SilcharSilchar · India6 works · 72 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksComputer Vision and Pattern Recognition · 2,869 works

Cite

Cite as L. D. M. S. Sai Teja, N. Siva Gopala Krishna, Ufaq Khan, Muhammad Haris Khan and Atul Mishra. 2026. DAMASHA: Detecting AI in Mixed Adversarial Texts via Segmentation with Human-interpretable Attribution. Findings of the Association for Computational Linguistics: EACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-eacl.326

@inproceedings{teja2026damasha,
  title = {{DAMASHA: Detecting AI in Mixed Adversarial Texts via Segmentation with Human-interpretable Attribution}},
  author = {L. D. M. S. Sai Teja, N. Siva Gopala Krishna, Ufaq Khan, Muhammad Haris Khan and Atul Mishra},
  booktitle = {Findings of the Association for Computational Linguistics: EACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-eacl.326},
  url = {https://doi.org/10.18653/v1/2026.findings-eacl.326}
}

References

46 references, 24 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Moustafa Alzantot, Yash Sharma, Ahmed Elgohary, Bo-Jhang Ho, Mani Srivastava, and Kai-Wei Chang. 2018. Generating natural language adversarial examples. arXiv preprint arXiv:1804.07998.
  2. Guangsheng Bao, Yanbin Zhao, Zhiyang Teng, Linyi Yang, and Yue Zhang. 2023. Fast-detectgpt: Efficient zero-shot detection of machine-generated text via conditional probability curvature. arXiv preprint arXiv:2310.05130.
  3. Chaka Chaka. 2023. Detecting ai content in responses generated by chatgpt, youchat, and chatsonic: The case of five ai content detection tools. Journal of Applied Learning and Teaching, 6(2):94–104. find
  4. Liam Dugan, Daphne Ippolito, Arun Kirubarajan, and Chris Callison-Burch. 2020. Roft: A tool for evaluating human detection of machine-generated text. arXiv preprint arXiv:2010.03070. find
  5. Javid Ebrahimi, Anyi Rao, Daniel Lowd, and Dejing Dou. 2018. HotFlip: White-box adversarial examples for text classification. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 31–36, Melbourne, Australia. Association for Comput…
  6. Ahmed M Elkhatat, Khaled Elsaid, and Saeed Almeer. 2023. Evaluating the efficacy of ai content detection tools in differentiating between human and ai-generated text. International Journal for Educational Integrity, 19(1):1–16. find
  7. Siddhant Garg and Goutham Ramakrishnan. 2020. Bae: Bert-based adversarial examples for text classification. arXiv preprint arXiv:2004.01970.
  8. Shreya Goyal, Sumanth Doddapaneni, Mitesh M Khapra, and Balaraman Ravindran. 2023. A survey of adversarial defenses and robustness in nlp. ACM Computing Surveys, 55(14s):1–39. find
  9. A Hans, A Schwarzschild, V Cherepanova, H Kazemi, A Saha, M Goldblum, J Geiping, and T Goldstein. 2024. Spotting llms with binoculars: Zero-shot detection of machine-generated text, 2024. URL: https://arxiv.org/abs/2401.12070. ↗
  10. Pengcheng He, Jianfeng Gao, and Weizhu Chen. 2021. Debertav3: Improving deberta using electra-style pre-training with gradient-disentangled embedding sharing. arXiv preprint arXiv:2111.09543.
  11. Guanhua Huang, Yuchen Zhang, Zhe Li, Yongjian You, Mingze Wang, and Zhouwang Yang. 2024. Are ai-generated text detectors robust to adversarial perturbations? arXiv preprint arXiv:2406.01179. find
  12. Zhiheng Huang, Wei Xu, and Kai Yu. 2015. Bidirectional lstm-crf models for sequence tagging. arXiv preprint arXiv:1508.01991.
  13. Daphne Ippolito, Daniel Duckworth, Chris Callison-Burch, and Douglas Eck. 2019. Automatic detection of generated text is easiest when humans are fooled. arXiv preprint arXiv:1911.00650.
  14. Ganesh Jawahar, Muhammad Abdul-Mageed, and Laks VS Lakshmanan. 2020. Automatic detection of machine generated text: A critical survey. arXiv preprint arXiv:2011.01314.
  15. Robin Jia and Percy Liang. 2017. Adversarial examples for evaluating reading comprehension systems. arXiv preprint arXiv:1707.07328.
  16. Di Jin, Zhijing Jin, Joey Tianyi Zhou, and Peter Szolovits. 2020. Is bert really robust? a strong baseline for natural language attack on text classification and entailment. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 8018–8025.
  17. Mandar Joshi, Danqi Chen, Yinhan Liu, Daniel S Weld, Luke Zettlemoyer, and Omer Levy. 2020. Spanbert: Improving pre-training by representing and predicting spans. Transactions of the association for computational linguistics, 8:64–77.
  18. Ahmed K Kadhim, Lei Jiao, Rishad Shafik, and Ole-Christoffer Granmo. 2025. Adversarial attacks on ai-generated text detection models: A token probability-based approach using embeddings. arXiv preprint arXiv:2501.18998. — [arXiv:2501.18998](https://arxiv.org/abs/2501.18998) find
  19. Omri Koshorek, Adir Cohen, Noam Mor, Michael Rotman, and Jonathan Berant. 2018. Text segmentation as a supervised learning task. arXiv preprint arXiv:1803.09337.
  20. Kalpesh Krishna, Yixiao Song, Marzena Karpinska, John Wieting, and Mohit Iyyer. 2023. Paraphrasing evades detectors of ai-generated text, but retrieval is an effective defense. Advances in Neural Information Processing Systems, 36:27469–27500.
  21. Laida Kushnareva, Tatiana Gaintseva, German Magai, Serguei Barannikov, Dmitry Abulkhanov, Kristian Kuznetsov, Eduard Tulchinskii, Irina Piontkovskaya, and Sergey Nikolenko. 2023. Ai-generated text boundary detection with roft. arXiv preprint arXiv:2311.08349. — [arXiv:2311.08349](https://arxiv.org/abs/2311.08349) find
  22. Mina Lee, Percy Liang, and Qian Yang. 2022. Coauthor: Designing a human-ai collaborative writing dataset for exploring language model capabilities. In Proceedings of the 2022 CHI conference on human factors in computing systems, pages 1–19. find
  23. Dianqi Li, Yizhe Zhang, Hao Peng, Liqun Chen, Chris Brockett, Ming-Ting Sun, and Bill Dolan. 2021. Contextualized perturbation for textual adversarial attack. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technolo…
  24. Jinfeng Li, Tianyu Du, Shouling Ji, Rong Zhang, Quan Lu, Min Yang, and Ting Wang. 2020. {TextShield}: robust text classification based on multimodal embedding and neural machine translation. In 29th USENIX Security Symposium (USENIX Security 20), pages 1381–1398. find
  25. Yafu Li, Qintong Li, Leyang Cui, Wei Bi, Zhilin Wang, Longyue Wang, Linyi Yang, Shuming Shi, and Yue Zhang. 2024. MAGE: Machine-generated text detection in the wild. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 36–53, Bangk…
  26. Yepeng Liu, Xuandong Zhao, Christopher Kruegel, Dawn Song, and Yuheng Bu. 2025. In-context watermarks for large language models. arXiv preprint arXiv:2505.16934. find
  27. 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. arXiv preprint arXiv:1907.11692.
  28. Elyas Masrour, Bradley Emi, and Max Spero. 2025. Damage: Detecting adversarially modified ai generated text. arXiv preprint arXiv:2501.03437. find
  29. Eric Mitchell, Yoonho Lee, Alexander Khazatsky, Christopher D Manning, and Chelsea Finn. 2023. Detectgpt: Zero-shot machine-generated text detection using probability curvature. In International conference on machine learning, pages 24950–24962. PMLR.
  30. John X Morris, Eli Lifland, Jin Yong Yoo, Jake Grigsby, Di Jin, and Yanjun Qi. 2020. Textattack: A framework for adversarial attacks, data augmentation, and adversarial training in nlp. arXiv preprint arXiv:2005.05909.
  31. Shashi Narayan, Shay B. Cohen, and Mirella Lapata. 2018. Don’t give me the details, just the summary! topic-aware convolutional neural networks for extreme summarization. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 1797–1807, Brussels, Belgium. As…
  32. Jonathon Read, Rebecca Dridan, Stephan Oepen, and Lars Jørgen Solberg. 2012. Sentence boundary detection: A long solved problem? In Proceedings of COLING 2012: Posters, pages 985–994.
  33. Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2018. Semantically equivalent adversarial rules for debugging nlp models. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (volume 1: long papers), pages 856–865. find
  34. Vinu Sankar Sadasivan, Aounon Kumar, Sriram Balasubramanian, Wenxiao Wang, and Soheil Feizi. 2023. find
  35. Can ai-generated text be reliably detected? arXiv preprint arXiv:2303.11156.
  36. Brian Tufts, Xuandong Zhao, and Lei Li. 2024. A practical examination of ai-generated text detectors for large language models. arXiv preprint arXiv:2412.05139. find
  37. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems, 30.
  38. Pengyu Wang, Linyang Li, Ke Ren, Botian Jiang, Dong Zhang, and Xipeng Qiu. 2023a. Seqxgpt: Sentence-level ai-generated text detection. arXiv preprint arXiv:2310.08903. find
  39. Yuxia Wang, Jonibek Mansurov, Petar Ivanov, Jinyan Su, Artem Shelmanov, Akim Tsvigun, Chenxi Whitehouse, Osama Mohammed Afzal, Tarek Mahmoud, Toru Sasaki, and 1 others. 2023b. M4: Multi-generator, multi-domain, and multi-lingual black-box machine-generated text detection. arXiv preprint arXiv:2305.1… find
  40. Benjamin Warner, Antoine Chaffin, Benjamin Clavié, Orion Weller, Oskar Hallström, Said Taghadouini, Alexis Gallagher, Raja Biswas, Faisal Ladhak, Tom Aarsen, and 1 others. 2024. Smarter, better, faster, longer: A modern bidirectional encoder for fast, memory efficient, and long context finetuning an…
  41. Junchao Wu, Shu Yang, Runzhe Zhan, Yulin Yuan, Lidia Sam Chao, and Derek Fai Wong. 2025. A survey on llm-generated text detection: Necessity, methods, and future directions. Computational Linguistics, 51(1):275–338.
  42. Zhuohan Xie, Trevor Cohn, and Jey Han Lau. 2023. The next chapter: A study of large language models in storytelling. arXiv preprint arXiv:2301.09790. find
  43. Xianjun Yang, Liangming Pan, Xuandong Zhao, Haifeng Chen, Linda Ruth Petzold, William Yang Wang, and Wei Cheng. 2024. A survey on detection of LLMs-generated content. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 9786–9805, Miami, Florida, USA. Association for Compu… find
  44. Zijie Zeng, Shiqi Liu, Lele Sha, Zhuang Li, Kaixun Yang, Sannyuya Liu, Dragan Gašević, and Guanliang Chen. 2024a. Detecting ai-generated sentences in human-ai collaborative hybrid texts: Challenges, strategies, and insights. arXiv preprint arXiv:2403.03506. — [arXiv:2403.03506](https://arxiv.org/abs/2403.03506) find
  45. Zijie Zeng, Lele Sha, Yuheng Li, Kaixun Yang, Dragan Gašević, and Guangliang Chen. 2024b. Towards automatic boundary detection for human-ai collaborative hybrid essay in education. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 22502–22510. find
  46. Qihui Zhang, Chujie Gao, Dongping Chen, Yue Huang, Yixin Huang, Zhenyang Sun, Shilin Zhang, Weiye Li, Zhengyan Fu, Yao Wan, and 1 others. 2024. Llm-as-a-coauthor: Can mixed human-written and machine-generated text be detected? arXiv preprint arXiv:2401.05952. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/teja-2026-damasha-detecting-ai-mixed-adversarial/