Pith. sign in

REVIEW 4 major objections 4 minor 67 references

A General Method for Detecting Information Generated by Large Language Models

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A detector trained on five LLMs flags unseen models at 0.888 AUC.

desk verdict Builds a novel detector architecture and the LOGO evaluation is clean, but the 'general' claim rests on a single generation mode—every test doc is a continuation of a human prefix. read the letter →

arxiv 2506.21589 v1 pith:5AS7BL7W submitted 2025-06-18 cs.CL

Minjia Mao , Dongjun Wei , Xiao Fang , Michael Chau This is my paper · ORCID
classification cs.CL
keywords largelanguagemodeldetectionmethodLLM-generatedinformationgenerativedeeplearningmemorynetworkdiscrepancymitigation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper introduces a detector meant to tell human-written text apart from text produced by a large language model even when the model and the topic were absent from training. It claims that a detector trained on five LLMs across five domains reaches an average AUC of 0.888 and F1 of 0.826 on unseen model-domain pairs, outperforming zero-shot, feature-based, and fine-tuning baselines. The two design ideas are twin memory networks, which learn separate author and domain embeddings, and a theory-guided detection generalization module that aligns embeddings across authors and domains. If the claim holds, platforms can flag machine-generated content without retraining for every new model or topic that appears.

What carries the argument

The load-bearing machinery is the pair of modules: Twin Memory Networks (TMN) and the Detection Generalization Module (DGM). TMN stores per-author and per-domain memory banks initialized by K-means and read by a two-level hierarchical attention network, producing document-specific author and domain embeddings that are concatenated with the textual embedding. DGM is built from Theorem 1, a generalization bound that expresses the error on unseen LLMs and domains as a weighted sum of training errors plus two maximal H-divergence terms, where H-divergence is a standard distribution-distance measure. DGM minimizes those terms with MMD-based discrepancy losses so that human embeddings become domain-invariant and LLM embeddings become both LLM- and domain-invariant. The theorem's constant C is treated as small, so the practical detector is the direct engineering of the bound.

What would settle it

Regenerate the evaluation with documents that LLMs produce from scratch, or as summaries, translations, or paraphrases, instead of continuing a human prefix, holding all models and domains fixed; if GLD's AUC drops toward the baselines or chance, the claimed generalization to unseen LLMs and domains is an artifact of the continuation protocol rather than a general signal.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that generalization to unseen LLMs and unseen domains is achievable by making the detector's internal embeddings invariant to which LLM wrote the text and which domain it came from. The paper derives a generalization bound (Theorem 1) on the error a detector will suffer on an unseen distribution, and the bound decomposes into training errors, the divergence between human-written embeddings across domains, and the divergence between LLM-generated embeddings across LLM-domain pairs. GLD realizes the bound by training a classifier on those terms, using Maximum Mean Discrepancy losses to pull embeddings closer together. Empirically, with leave-one-group-out evaluation over 25 unseen LLM-domain combinations, GLD reports AUC 0.888 and F1 0.826 with statistically significant improvements over all benchmarks.

Load-bearing premise

The empirical claim rests on every LLM document being created by prompting the model to continue the first 30 words of a human-written document, and the theoretical guarantee additionally assumes the unseen distribution sits close to the training distributions, so if real-world machine text is written from scratch or the unseen style drifts far away, the claimed generalization may be an artifact of that setup rather than a general property of LLM authorship.

Editorial extensions

If this is right

  • A single detector could monitor social feeds and review sites for machine-generated content without being updated for every new model release or niche topic.
  • Because GLD needs author and domain labels only at training time, it can be deployed at test time without knowing which model wrote a document or which topic it concerns.
  • Platform operators could use GLD to flag or label AI-written posts, supporting regulations that require disclosure of LLM-generated content.
  • The reported stability, with the lowest standard deviations across the 25 unseen group pairs, implies the detector is less biased toward particular models or topics than the baselines.
  • Since the method builds on a transformer encoder backbone, improvements in that encoder should directly raise detection accuracy.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If real-world LLM text is usually free-form, summarized, or rewritten rather than continued from a human prefix, the reported gains may shrink; this is testable by regenerating the evaluation under those generation modes.
  • The same Theorem 1 structure could be reused for other author/domain pairs, such as detecting machine-generated code or structured data, whenever author and domain labels are available for training.
  • The 30-word continuation protocol may mean the detector is partly reading the effect of a model continuing a human's style; an ablation comparing continuation from human-written versus machine-written prefixes would isolate that.
  • The paper leaves the magnitude of the theorem's constant C unmeasured; checking whether unseen distributions stay close to the training convex hull would separate the empirical success from the theoretical guarantee.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. This paper proposes GLD, a learning-based detector for LLM-generated text that combines twin memory networks (for author and domain embeddings) with a detection generalization module that adds MMD-based discrepancy losses to a classifier loss. The authors derive an upper bound on the error for unseen LLM/domain distributions (Theorem 1, Appendix A) and use it to justify the loss design. They evaluate GLD under leave-one-group-out across five LLMs and five domains, report average AUC 0.888 and F1 0.826, compare with zero-shot, feature-based, and fine-tuning baselines, run ablations, and provide two case studies with Gemini 2.0 Pro on reviews and academic conclusions.

Significance. If the empirical and theoretical claims hold, the paper would be a useful contribution to machine-generated text detection: it provides a systematic LOGO evaluation over 25 unseen LLM-domain pairs, consistent gains over strong baselines, ablation evidence for the main design components, and an explicit generalization bound connecting domain-adaptation theory to the detector design. The reproducible implementation details in Appendix B and the inclusion of both zero-shot and learning-based baselines are strengths. However, the reported generalization is currently demonstrated only for one text-generation protocol, and the link between Theorem 1 and the actual losses has unproved steps (MMD treated as the empirical H-divergence; unweighted loss used in place of the pi*-weighted source error), so the significance is conditional on additional evidence and theoretical tightening.

major comments (4)
  1. [Section 5.1, Section 5.5] The central claim of generality is not supported beyond a single generation mode. In Section 5.1, every LLM-generated training and test document is produced by prompting the LLM to continue from the first 30 words of a human-written document, and the two case studies in Section 5.5 use the same continuation protocol (reviews) or conditioning on human title/abstract (academic conclusions). The leave-one-group-out results in Table 5 therefore establish generalization across unseen LLMs and domains only within this prefix-continuation mode. A detector could be exploiting the distributional seam at the human/LLM boundary rather than a general property of LLM authorship; free-form, summarization, and rewriting are untested. Please add experiments that generate positives without a human prefix (e.g., free-form generation from a topic, summarization, paraphrasing) and report LOGO results for each mode; alternatively, explicitly scope the claims to continuation-based generation.
  2. [Section 4.2.1, Eqs. (10)-(13)] The paper equates the empirical H-divergence d_hat_H with MMD without proof. Equation (10) defines MMD with a multi-Gaussian kernel, and the text states that the empirical H-divergence is 'measured using' MMD, then replaces max d_hat_H in Theorem 1 by max MMD in Eqs. (12)-(13). MMD is an integral probability metric over a fixed RKHS, whereas H-divergence is a supremum over a hypothesis class; the two are not equal in general and the manuscript gives no bound relating them in this setting. As a result L_h and L_g are not shown to minimize the second and third terms of Inequality (9). Please supply a proof or a known-result citation relating the MMD losses to the H-divergence terms, or present DGM as a heuristic domain-invariance regularizer rather than as an implementation of Theorem 1.
  3. [Section 4.2.2, Eq. (15) and Theorem 1] The training loss does not minimize the first term of the bound. Inequality (9) and Eq. (A3) in Appendix A involve the pi*-weighted sum of source errors, where pi* is the optimal mixture minimizing gamma = d_H(D_u, sum_ij pi_ij D_ij). The implemented classification loss L_y in Eq. (15) is an unweighted cross-entropy over all training documents, and the paper never estimates or uses pi*. Without an argument that uniform weighting approximates the pi* weighting, the first term of the bound is not actually minimized by the proposed objective. Please justify the uniform-weight choice theoretically or modify the objective to reflect pi*.
  4. [Appendix A, Eq. (A18)] The constant C in Theorem 1 contains gamma/2 and epsilon*, where gamma = min_pi d_H(D_u, sum_ij pi_ij D_ij) measures how far the unseen distribution is from the convex hull of training distributions. No empirical check or theoretical bound on gamma is provided, so the bound may be vacuous for the unseen domains tested in Section 5.3. Please report an estimate or upper bound of gamma (or discuss why C can be treated as small) so that the theory substantiates rather than merely accompanies the empirical results.
minor comments (4)
  1. [Eq. (4)] The indexing in the attention weight list appears to be a typo: it reads [b0 b2 ... bm] but should presumably be [b0 b1 ... bm].
  2. [Appendix A, Eq. (A15)] In Eq. (A15), the right-hand side's first H-divergence term appears to be missing the empirical hat: the displayed 'dH' should likely be 'd_hat_H' to match the conclusion of Lemma 4.
  3. [Section 5.2 and Table 5] The paper reports that GLD's improvements are statistically significant at the 5% level but does not state which test was used, whether it is paired across the 25 LOGO folds, or whether any multiple-comparison correction was applied.
  4. [Section 5.1, footnote 13] The claim that truncation lengths of 20, 30, and 40 words give statistically indistinguishable results is not supported with any numbers; please provide the comparison or remove the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: Theorem 1 is imported transparently from external prior work (Albuquerque et al. 2019), the DGM losses minimize that bound's terms, and no fitted result is relabeled as a prediction.

full rationale

GLD's design is guided by Theorem 1, which upper-bounds the error on an unseen distribution D_u by a convex combination of per-group training errors, the maximal pairwise H-divergences among human and LLM-generated embedding distributions, and a constant C. The bound is not derived from the method's own outputs or from results unique to the authors. Appendix A states: 'Following Wang et al. (2022), we restate Theorem 1 from Albuquerque et al. (2019) in the context of our study as Lemma 1.' Albuquerque et al. (2019) is external prior work with no author overlap, so this is independent support rather than a self-citation chain. The DGM losses (Equations 12, 13, and 15) are constructed to minimize the terms of this externally sourced bound; this is theory-guided design, not a self-referential prediction. The reported LOGO AUC and F1 scores (Table 5) come from held-out LLM-domain groups under a leave-one-group-out protocol, and no fitted test-set constant is presented as a prediction; the hyperparameters in Appendix B are fixed training hyperparameters, not post hoc fits to the evaluation outcome. The paper's self-citations (e.g., Fang et al. 2024 for the prevalence of LLM-generated misinformation) are motivational and carry no load in the derivation or in the evaluation. The substitution of MMD for empirical H-divergence in Equations 10-13 is an unsupported approximation and a potential correctness risk, but it is not circularity because the losses are not defined in terms of the measured outcome and no theorem or result is invoked that presupposes the effectiveness of GLD. Therefore, under the required standard of exhibiting a specific reduction of a prediction to its inputs by construction, no circular step can be identified.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The method depends on several explicitly chosen hyperparameters (Q, tau, beta, lambda weights, bandwidths) and on assumptions imported from domain adaptation theory. The most fragile assumption is that MMD can replace H-divergence in the bound, which is not established. No new physical or conceptual entities are introduced.

free parameters (6)
  • Number of memory units Q = 10
    Chosen by hand; controls granularity of author and domain prototypes in memory banks.
  • Attention temperature tau
    Hyperparameter in Equations 2 and 4; no value given, affects smoothness of attention weights.
  • Memory update strength beta
    Hyperparameter in Equation 6; no value given, controls how much memory units change per document.
  • Loss weights lambda_y, lambda_h, lambda_g = 0.1, 0.2, 0.2
    Weights in Equation 16, set in Appendix B.
  • MMD bandwidth set R = {2^-3, ..., 2^1}
    Bandwidths for the multi-Gaussian kernel in Equation 11, set in Appendix B.
  • Learning rate and epochs = 5e-5, 4 epochs
    Training schedule in Appendix B; not justified by validation.
assumptions (4)
  • domain assumption The H-divergence generalization bound of Albuquerque et al. (2019), restated as Lemma 1, is valid in this setting.
    The paper leans on this prior bound to derive Theorem 1; if the assumptions of that bound (e.g., hypothesis space and distribution convex hull) do not hold, the theoretical motivation for the loss design collapses.
  • domain assumption The unseen distribution D_u is close to the convex hull of training distributions, so gamma in Equation A18 is small.
    Theorem 1's constant C contains gamma, which is unknown; the paper asserts C is small without evidence. The whole promise of generalization depends on D_u not being far from the training mixtures.
  • ad hoc to paper MMD with a multi-Gaussian kernel can stand in for the empirical H-divergence d_hat_H.
    Section 4.2.1 equates these two quantities, but MMD in an RKHS is not the H-divergence defined in Appendix A; this substitution is not proven, so the losses L_h and L_g are not rigorously tied to Theorem 1.
  • ad hoc to paper A uniformly weighted classification loss approximates the pi*-weighted source error in Theorem 1.
    Theorem 1 weights source errors by unknown optimal pi*, but the training loss L_y in Equation 15 weights every document equally; no justification is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A General Method for Detecting Information Generated by Large Language Models." pith.science (2026). https://pith.science/paper/5AS7BL7W

@misc{pith2026250621589,
  author       = {Pith},
  title        = {Pith review of: A General Method for Detecting Information Generated by Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5AS7BL7W}},
  note         = {Machine review of arXiv:2506.21589}
}
read the original abstract

The proliferation of large language models (LLMs) has significantly transformed the digital information landscape, making it increasingly challenging to distinguish between human-written and LLM-generated content. Detecting LLM-generated information is essential for preserving trust on digital platforms (e.g., social media and e-commerce sites) and preventing the spread of misinformation, a topic that has garnered significant attention in IS research. However, current detection methods, which primarily focus on identifying content generated by specific LLMs in known domains, face challenges in generalizing to new (i.e., unseen) LLMs and domains. This limitation reduces their effectiveness in real-world applications, where the number of LLMs is rapidly multiplying and content spans a vast array of domains. In response, we introduce a general LLM detector (GLD) that combines a twin memory networks design and a theory-guided detection generalization module to detect LLM-generated information across unseen LLMs and domains. Using real-world datasets, we conduct extensive empirical evaluations and case studies to demonstrate the superiority of GLD over state-of-the-art detection methods. The study has important academic and practical implications for digital platforms and LLMs.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 49 canonical work pages

  1. [1]

    H., and Mitliagkas, I

    Albuquerque, I., Monteiro, J., Darvishi, M., Falk, T. H., and Mitliagkas, I. (2019). Generalizing to unseen domains via distribution matching. arXiv preprint arXiv:1911.00804

    arXiv 2019

  2. [2]

    Ben-David, S., Blitzer, J., Crammer, K., Kulesza, A., Pereira, F., and Vaughan, J. W. (2010). A theory of learning from different domains. Machine Learning , 79:151--175

    work page 2010

  3. [3]

    Kifer, D., Ben-David, S., and Gehrke, J. (2004). Detecting change in data streams. In Proceedings of the Thirtieth International Conference on Very Large Data Bases , volume 30, pages 180--191

    work page 2004

  4. [4]

    Kingma, D. P. and Ba, J. (2015). Adam: A method for stochastic optimization. In Proceedings of the 3rd International Conference on Learning Representations

    2015

  5. [5]

    Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., and Liu, P. J. (2020). Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research , 21(140):1--67

    2020

  6. [6]

    Sanh, V. (2019). Distilbert, a distilled version of bert: Smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108

    arXiv 2019

  7. [7]

    Wang, J., Lan, C., Liu, C., Ouyang, Y., Qin, T., Lu, W., Chen, Y., Zeng, W., and Philip, S. Y. (2022). Generalizing to unseen domains: A survey on domain generalization. IEEE Transactions on Knowledge and Data Engineering , 35(8):8052--8072

    work page 2022

  8. [8]

    , " * write output.state after.block = add.period write newline

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution isbn issn journal key month note number organization pages publisher school series title type url volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1...

Show all 67 references
  1. [9]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  2. [10]

    Abbasi, A., Zhang, Z., Zimbra, D., Chen, H., and Nunamaker Jr, J. F. (2010). Detecting fake websites: The contribution of statistical learning theory. MIS Quarterly , pages 435--461

    2010

  3. [11]

    Bao, G., Zhao, Y., Teng, Z., Yang, L., and Zhang, Y. (2023). Fast-detectgpt: Efficient zero-shot detection of machine-generated text via conditional probability curvature. In The Twelfth International Conference on Learning Representations

    2023

  4. [12]

    E., and Cohan, A

    Beltagy, I., Peters, M. E., and Cohan, A. (2020). Longformer: The long-document transformer. arXiv preprint arXiv:2004.05150

    2020 arXiv

  5. [13]

    Berente, N., Gu, B., Recker, J., and Santhanam, R. (2021). Managing artificial intelligence. MIS Quarterly , 45(3)

    2021

  6. [14]

    Bhattacharjee, A., Moraffah, R., Garland, J., and Liu, H. (2024). Eagle: A domain generalization framework for ai-generated text detection. arXiv preprint arXiv:2403.15690

    2024 arXiv

  7. [15]

    Black, S., Biderman, S., Hallahan, E., Anthony, Q., Gao, L., Golding, L., He, H., Leahy, C., McDonell, K., Phang, J., et al. (2022). Gpt-neox-20b: An open-source autoregressive language model. In Proceedings of BigScience Episode\# 5--Workshop on Challenges & Perspectives in C...

    2022

  8. [16]

    M., Gretton, A., Rasch, M

    Borgwardt, K. M., Gretton, A., Rasch, M. J., Kriegel, H.-P., Sch \"o lkopf, B., and Smola, A. J. (2006). Integrating structured biological data by kernel maximum mean discrepancy. Bioinformatics , 22(14):e49--e57

    2006

  9. [17]

    Church, K. W. (2017). Word2vec. Natural Language Engineering , 23(1):155--162

    2017

  10. [18]

    Devlin, J., Chang, M.-W., Lee, K., and Toutanova, K. (2019). Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics , pages 4171--4186

    2019

  11. [19]

    Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Letman, A., Mathur, A., Schelten, A., Yang, A., Fan, A., et al. (2024). The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  12. [20]

    Else, H. (2023). Abstracts written by chatgpt fool scientists. Nature , 613:423

    2023

  13. [21]

    Fan, A., Jernite, Y., Perez, E., Grangier, D., Weston, J., and Auli, M. (2019). Eli5: Long form question answering. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics , pages 3558--3567

    2019

  14. [22]

    Fang, X., Che, S., Mao, M., Zhang, H., Zhao, M., and Zhao, X. (2024). Bias of ai-generated content: an examination of news produced by large language models. Scientific Reports , 14(1):1--20

    2024

  15. [23]

    and Zubiaga, A

    Fr \"o hling, L. and Zubiaga, A. (2021). Feature-based detection of automated language models: tackling gpt-2, gpt-3 and grover. PeerJ Computer Science , 7:e443

    2021

  16. [24]

    Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., March, M., and Lempitsky, V. (2016). Domain-adversarial training of neural networks. Journal of Machine Learning Research , 17(59):1--35

    2016

  17. [25]

    Gehrmann, S., Strobelt, H., and Rush, A. M. (2019). Gltr: Statistical detection and visualization of generated text. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations , pages 111--116

    2019

  18. [26]

    W., Huang, T., Socrates, V., Chi, L., Taylor, R

    Gilson, A., Safranek, C. W., Huang, T., Socrates, V., Chi, L., Taylor, R. A., Chartash, D., et al. (2023). How does chatgpt perform on the united states medical licensing examination (usmle)? the implications of large language models for medical education and knowledge assessm...

    2023

  19. [27]

    and Singh, A

    Goli, A. and Singh, A. (2024). Frontiers: Can large language models capture human preferences? Marketing Science , 43(4):709--722

    2024

  20. [28]

    Graves, A., Wayne, G., and Danihelka, I. (2014). Neural turing machines. arXiv preprint arXiv:1410.5401

    2014 arXiv

  21. [29]

    Guo, B., Zhang, X., Wang, Z., Jiang, M., Nie, J., Ding, Y., Yue, J., and Wu, Y. (2023). How close is chatgpt to human experts? comparison corpus, evaluation, and detection. arXiv preprint arXiv:2301.07597

    2023 arXiv

  22. [30]

    Guo, Y., Shang, G., and Clavel, C. (2024a). Benchmarking linguistic diversity of large language models. arXiv preprint arXiv:2412.10271

    2024 arXiv

  23. [31]

    Guo, Z., Jiao, K., Yao, X., Wan, Y., Li, H., Xu, B., Zhang, L., Wang, Q., Zhang, Y., and Mao, Z. (2024b). Ustc-bupt at semeval-2024 task 8: Enhancing machine-generated text detection via domain adversarial neural networks and llm embeddings. In Proceedings of the 18th Internat...

    2024

  24. [32]

    Gupta, R., Jindal, V., and Kashyap, I. (2024). Recent state-of-the-art of fake review detection: a comprehensive review. The Knowledge Engineering Review , 39:e8

    2024

  25. [33]

    He, S., Hollenbeck, B., and Proserpio, D. (2022). The market for fake reviews. Marketing Science , 41(5):896--921

    2022

  26. [34]

    T., and Naaman, M

    Jakesch, M., Hancock, J. T., and Naaman, M. (2023). Human heuristics for ai-generated language are flawed. Proceedings of the National Academy of Sciences , 120(11):e2208839120

    2023

  27. [35]

    Jin, Q., Dhingra, B., Liu, Z., Cohen, W., and Lu, X. (2019). Pubmedqa: A dataset for biomedical research question answering. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language...

    2019

  28. [36]

    R., Srinivasan, K., Kong, S

    Kim, B. R., Srinivasan, K., Kong, S. H., Kim, J. H., Shin, C. S., and Ram, S. (2023). Rolex: A novel method for interpretable machine learning using robust local explanations. MIS Quarterly , 47(3)

    2023

  29. [37]

    Kirchenbauer, J., Geiping, J., Wen, Y., Katz, J., Miers, I., and Goldstein, T. (2023). A watermark for large language models. In International Conference on Machine Learning , pages 17061--17084. PMLR

    2023

  30. [38]

    Lappas, T., Sabnis, G., and Valkanas, G. (2016). The impact of fake reviews on online visibility: A vulnerability assessment of the hotel industry. Information Systems Research , 27(4):940--961

    2016

  31. [39]

    H., Han, S

    Lee, S., Shin, D., Kwon, K. H., Han, S. P., and Lee, S. K. (2024). Disinformation spillover: Uncovering the ripple effect of bot-assisted fake social engagement on public attention. MIS Quarterly , 48(3)

    2024

  32. [40]

    Li, Y., Li, Q., Cui, L., Bi, W., Wang, Z., Wang, L., Yang, L., Shi, S., and Zhang, Y. (2023). Mage: Machine-generated text detection in the wild. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics , pages 36--53

    2023

  33. [41]

    Liu, A., Pan, L., Lu, Y., Li, J., Hu, X., Zhang, X., Wen, L., King, I., Xiong, H., and Yu, P. (2024). A survey of text watermarking in the era of large language models. ACM Computing Surveys , 57(2):1--36

    2024

  34. [42]

    Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., and Stoyanov, V. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692

    2019 arXiv

  35. [43]

    Liu, Y., Zhang, Z., Zhang, W., Yue, S., Zhao, X., Cheng, X., Zhang, Y., and Hu, H. (2023). Argugpt: evaluating, understanding and identifying argumentative essays generated by gpt models. arXiv preprint arXiv:2304.07666

    2023 arXiv

  36. [44]

    Liu, Z., Xu, Y., Yu, T., Dai, W., Ji, Z., Cahyawijaya, S., Madotto, A., and Fung, P. (2021). Crossner: Evaluating cross-domain named entity recognition. In Proceedings of the AAAI Conference on Artificial Intelligence , pages 13452--13460

    2021

  37. [45]

    Long, M., Cao, Y., Wang, J., and Jordan, M. (2015). Learning transferable features with deep adaptation networks. In Proceedings of the 32nd International Conference on Machine Learning , pages 97--105. PMLR

    2015

  38. [46]

    D., and Finn, C

    Mitchell, E., Lee, Y., Khazatsky, A., Manning, C. D., and Finn, C. (2023). Detectgpt: Zero-shot machine-generated text detection using probability curvature. In International Conference on Machine Learning , pages 24950--24962. PMLR

    2023

  39. [47]

    L., Minas, R

    Moravec, P. L., Minas, R. K., and Dennis, A. R. (2019). Fake news on social media: People believe what they want to believe when it makes no sense at all. MIS Quarterly , 43(4)

    2019

  40. [48]

    Mostafazadeh, N., Chambers, N., He, X., Parikh, D., Batra, D., Vanderwende, L., Kohli, P., and Allen, J. (2016). A corpus and cloze evaluation for deeper understanding of commonsense stories. In Proceedings of the 2016 Conference of the North American Chapter of the Associatio...

    2016

  41. [49]

    Mousavi, R., Raghu, T., and Frey, K. (2020). Harnessing artificial intelligence to improve the quality of answers in online question-answering health forums. Journal of Management Information Systems , 37(4):1073--1098

    2020

  42. [50]

    Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022). Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems , 35:27730--27744

    2022

  43. [51]

    Padmanabhan, B., Fang, X., Sahoo, N., and Burton-Jones, A. (2022). Machine learning in information systems research. MIS Quarterly , 46(1)

    2022

  44. [52]

    and Rand, D

    Pennycook, G. and Rand, D. G. (2021). The psychology of fake news. Trends in Cognitive Sciences , 25(5):388--402

    2021

  45. [53]

    R., Ketter, W., Parsons, J., Rao, H

    Rai, A., Burton-Jones, A., Chen, H., Gupta, A., Hevner, A. R., Ketter, W., Parsons, J., Rao, H. R., Sarkar, S., and Yoo, Y. (2017). Editor’s comments: Diversity of design science research. MIS Quarterly , 41(1):iii--xviii

    2017

  46. [54]

    Rajpurkar, P., Zhang, J., Lopyrev, K., and Liang, P. (2016). Squad: 100,000+ questions for machine comprehension of text. In Conference on Empirical Methods in Natural Language Processing , pages 2383--2392

    2016

  47. [55]

    A., and Dan, D

    Reisenbichler, M., Reutterer, T., Schweidel, D. A., and Dan, D. (2022). Frontiers: Supporting content marketing with natural language generation. Marketing Science , 41(3):441--452

    2022

  48. [56]

    Samtani, S., Chai, Y., and Chen, H. (2022). Linking exploits from the dark web to known vulnerabilities for proactive cyber threat intelligence: An attention-based deep structured semantic model. MIS Quarterly , 46(2)

    2022

  49. [57]

    W., Kreps, S., et al

    Solaiman, I., Brundage, M., Clark, J., Askell, A., Herbert-Voss, A., Wu, J., Radford, A., Krueger, G., Kim, J. W., Kreps, S., et al. (2019). Release strategies and the social impacts of language models. arXiv preprint arXiv:1908.09203

    2019 arXiv

  50. [58]

    and Van Noorden, R

    Stokel-Walker, C. and Van Noorden, R. (2023). What chatgpt and generative ai mean for science. Nature , 614(7947):214--216

    2023

  51. [59]

    Y., Wang, D., and Nakov, P

    Su, J., Zhuo, T. Y., Wang, D., and Nakov, P. (2023). Detectllm: Leveraging log rank information for zero-shot detection of machine-generated text. In The 2023 Conference on Empirical Methods in Natural Language Processing , pages 12395--12412

    2023

  52. [60]

    and Li, B

    Wang, W. and Li, B. (2024). Learning personalized privacy preference from public data. Information Systems Research

    2024

  53. [61]

    Wei, X., Zhang, Z., Zhang, M., Chen, W., and Zeng, D. D. (2022). Combining crowd and machine intelligence to detect false news on social media. MIS Quarterly , 46(2):977--1008

    2022

  54. [62]

    F., and Chao, L

    Wu, J., Yang, S., Zhan, R., Yuan, Y., Wong, D. F., and Chao, L. S. (2025). A survey on llm-gernerated text detection: Necessity, methods, and future directions. Computational Linguistics , pages 1--66

    2025

  55. [63]

    Zhang, D., Zhou, L., Tao, J., Zhu, T., and Gao, G. (2025). Ketch: a knowledge-enhanced transformer-based approach to suicidal ideation detection from social media content. Information Systems Research , 36(1):572--599

    2025

  56. [64]

    V., et al

    Zhang, S., Roller, S., Goyal, N., Artetxe, M., Chen, M., Chen, S., Dewan, C., Diab, M., Li, X., Lin, X. V., et al. (2022). Opt: Open pre-trained transformer language models. arXiv preprint arXiv:2205.01068

    2022 arXiv

  57. [65]

    Zhang, X., Zhao, J., and LeCun, Y. (2015). Character-level convolutional networks for text classification. Advances in Neural Information Processing Systems , 28

    2015

  58. [66]

    Zhao, X., Fang, X., He, J., and Huang, L. (2023). Exploiting expert knowledge for assigning firms to industries: A novel deep learning method. MIS Quarterly , 47(3):1147--1176

    2023

  59. [67]

    Zhou, J., Pang, Z., and Wang, Y.-X. (2024). Rmem: Restricted memory banks improve video object segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18602--18611

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.