Survey of Hallucination in Natural Language Generation
Open accessThe DOI matches its Crossref record; 355 of the 360 references it cites were found in a registry or an index, and 4 are not checked yet; 224 of the 256 sources its claims cite were checked for support.
Abstract
Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based language models. This advancement has led to more fluent and coherent NLG, leading to improved development in downstream tasks such as abstractive summarization, dialogue generation, and data-to-text generation. However, it is also apparent that deep learning based generation is prone to hallucinate unintended text, which degrades the system performance and fails to meet user expectations in many real-world scenarios. To address this issue, many studies have been presented in measuring and mitigating hallucinated texts, but these have never been reviewed in a comprehensive manner before. In this survey, we thus provide a broad overview of the research progress and challenges in the hallucination problem in NLG. The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions, and (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, and machine translation. This survey serves to facilitate collaborative efforts among researchers in tackling the challenge of hallucinated texts in NLG.
Verified 2026-11-19 · 4 references not checked yet
Overview
CoChat Summary
TL;DR
This survey provides a broad overview of hallucination in natural language generation, covering definitions, contributors, metrics, mitigation methods, and task-specific progress across summarization, dialogue, GQA, data-to-text, machine translation, vision-language generation, and LLMs. It distinguishes intrinsic from extrinsic hallucination and argues that hallucination must be treated differently across tasks and sources.
Key findings
- The survey categorizes hallucination into intrinsic hallucination, where generated output contradicts the source, and extrinsic hallucination, where output cannot be verified from the source; extrinsic hallucination is not always erroneous because it may come from factually correct external information.
- It reports prevalence examples from cited work: 25% of summaries generated by state-of-the-art abstractive summarization systems have hallucinated content; 62% of first sentences in WIKIBIO have additional information not stated in the corresponding infobox; and 74% of samples in the QMSum dataset consist of inconsistent facts.
- It reviews hallucination metrics as statistical, model-based (IE-based, QA-based, NLI-based, faithfulness classification, LM-based), and human evaluation, and mitigation methods as data-related and modeling/inference methods including architecture changes, planning/sketching, RL, multi-task learning, controllable generation, and post-processing.
- For LLMs, it notes that RLHF reduced the hallucination rate from 41% to 21% in closed-domain tasks for InstructGPT, and discusses reference-dependent and reference-free hallucination metrics plus mitigation via data refinement, RAG, safety fine-tuning, model editing, decoding methods, chain-of-thought variants, post-processing, and ensembling.
Methods
- Literature survey design: the paper reviews existing research on hallucination in NLG rather than conducting new experiments or primary data collection.
- Coverage: general NLG hallucination topics (definitions, contributors, metrics, mitigation, future directions) plus task-specific sections on abstractive summarization, dialogue generation, generative question answering, data-to-text generation, neural machine translation, vision-language generation, and large language models.
- Analysis: synthesizes and compares definitions, metrics, mitigation approaches, and future directions across tasks, with examples in tables and figures; the LLM section was updated in Jan 2024.
Limitations
- The paper states that GQA has no standard definition or categorization of hallucination and no automatic metric specifically for hallucination in GQA; it also notes that open-domain dialogue lacks a standard hallucination metric and that VL hallucination research is still in its early stage.
- The paper states that extrinsic hallucination has been somewhat overlooked and is more challenging to reduce, and that most existing mitigation methods do not distinguish between intrinsic and extrinsic hallucination.
- Evident from the text: as a survey, it does not provide new empirical validation of metrics or mitigation methods, and its LLM coverage is explicitly updated only through Jan 2024, so later work is not included.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
124 Valid · 26 Not valid · 19 Not checked · 135 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“Furthermore, perturbation, where the insertions of most common tokens are placed at the beginning of the sentence, seems to be the most successful in hallucination mitigation.”
What the sources say
- Not checked — no full text of the sourcecites [133] [133] Katherine Lee, Orhan Firat, Ashish Agarwal, Clara Fannjiang, and David Sussillo. 2019. Hallucinations in Neural Machine Translation. ICLR (2019). -
- Disputedcites [215] [215] Vikas Raunak, Arul Menezes, and Marcin Junczys-Dowmunt. 2021. The Curious Case of Hallucinations in Neural Machine Translation. (4 2021), 1172-1183. https://arxiv.org/abs/2104.06683v1 - sourceExperimental — accuracy not yet measured
In the passage describing the perturbation procedure (page unavailable), the paper says that inserting common tokens at the beginning is used to generate hallucinations, not mitigate them. Its prevention section instead discusses approaches such as robust learning and corpus-level filtering. (LLM, conf=0.92)
Checked by a reasoning model — may be wrong.
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“While there are works on the visual extension of faithfulness-oriented RLHF [303], these essentially involve labeling more data, which is expensive and difficult to scale.”
What the sources say
- Disputedcites [303] [303] Tianyu Yu, Yuan Yao, Haoye Zhang, Taiwen He, Yifeng Han, Ganqu Cui, Jinyi Hu, Zhiyuan Liu, Hai-Tao Zheng, Maosong Sun, and Tat-Seng Chua. 2023. RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedback. CoRR abs/2312.00849 (2023). https://doi.org/10.48550/ARXIV.2312.00849 arXiv:2312.00849 - sourceExperimental — accuracy not yet measured
RLHF-V does use human annotation for visual-model alignment, but it makes the contrary point about needing more labeled data. The Fine-grained Correctional Human Preference Collection section says its corrections improve data efficiency, and Section 4.3 reports comparable hallucination rates using only 200 preference examples versus an order of magnitude more ranking-labeled examples. Our copy does not establish a separate claim about annotation cost or scalability. (LLM, conf=0.88)
Checked by a reasoning model — may be wrong.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Disputed
Every source the paper cites for this claim was checked and supports it.
In the paper
“Nevertheless, as illustrated by Falke et al. [66], off-the-shelf NLI models tend to transfer poorly to the abstractive summarization task.”
What the sources say
- Supportedcites [66] [66] Tobias Falke, Leonardo FR Ribeiro, Prasetya Ajie Utama, Ido Dagan, and Iryna Gurevych. 2019. Ranking Generated Summaries by Correctness: An Interesting but Challenging Application for Natural Language Inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics . 2214-2220. - source
The Introduction states that models trained on NLI datasets transfer poorly to the paper’s downstream task of ranking generated summaries by correctness. The Conclusions likewise say that out-of-the-box NLI models do not perform well on that task. (LLM, conf=0.94)
- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Such wrong association results in generation with facts mixed up between two similar entities [60, 233].”
What the sources say
- Partly supportedcites [60] [60] Nouha Dziri, Andrea Madotto, Osmar Zaiane, and Avishek Joey Bose. 2021. Neural Path Hunter: Reducing Hallucination in Dialogue Systems via Path Grounding. EMNLP (2021). - sourceExperimental — accuracy not yet measured
The GPT2-KG passage describes an incorrect association between entities, and G Error Analysis gives an example of a generated response assigning a book to the wrong author. These support the claim that facts can be mixed up, but the paper does not say the entities involved are similar. The supplied passages do not provide page numbers. (LLM, conf=0.83)
- Supportedcites [233] [233] Kurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela, and Jason Weston. 2021. Retrieval Augmentation Reduces Hallucination in Conversation. EMNLP (2021). - source
The Introduction states that language models often mix up facts between two similar entities when generating factually incorrect statements. This directly supports the cited description of the resulting error. (LLM, conf=0.92)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“As a result, the target reference may contain information that cannot be supported by the source [198, 268].”
What the sources say
- Supportedcites [198] [198] Ankur Parikh, Xuezhi Wang, Sebastian Gehrmann, Manaal Faruqui, Bhuwan Dhingra, Diyi Yang, and Dipanjan Das. 2020. ToTTo: A Controlled Table-To-Text Generation Dataset. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) . 1173-1186. - source
The Dataset section says that reference targets selected from web sentences may contain information that cannot be inferred from the source. The Data Annotation Process section also describes deleting phrases unsupported by table cells. (LLM, conf=0.93)
- Supportedcites [268] [268] Hongmin Wang. 2019. Revisiting Challenges in Data-to-Text Generation with Fact Grounding. In Proceedings of the 12th International Conference on Natural Language Generation . 311-322. - source
The Introduction reports that about 40% of human-written game-summary content cannot be directly mapped to the input boxscore records and draws on information beyond those inputs. Section 2.1 likewise describes reference content not licensed by the boxscore table. This supports the claim that a target reference may contain information the source does not support. (LLM, conf=0.93)
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
117 more claims · 74 with their cited sources checked
Create a free account to see all 304 claims and to verify the 2 references still open.
- Borderline: one of the tests for a substantial claim is unclear. 1
- It may state two separate claims as one. 90
- Probably not a substantial claim. 142
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Hallucination can lead to serious consequences such as incorrect or misleading code, security risks, and ethical issues [186].”
From A survey on large language model based autonomous agents (2024) · cites this paper as [186]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“citation fabrication constitutes a form of extrinsic hallucination in the absence of retrieval augmentation, since the model generates references not grounded in any provided source material [4].”
From How LLMs Cite and Why It Matters: A Cross-Model Audit of Reference Fabrication in AI-Assisted Academic Writing and Methods to Detect Phantom Citations (2026) · cites this paper as [4]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Recent taxonomic work distinguishes at least seven cases [15,18].”
From Hallucination as epistemic pollution: epistemic injustice and credibility in futures-oriented science education (2026) · cites this paper as [18]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“LLMs struggle to generate accurate and up-to-date responses on current topics, due to outdated knowledge, lack of domain-specific expertise, or hallucination13–17.”
From Leveraging long context in retrieval augmented language models for medical question answering (2025) · cites this paper as [16]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Fine-tuning can enrich LLMs with domain knowledge, enhance their specific ability, improve the fairness and reliability of the outputs, and prevent certain damage caused by hallucination [26].”
From FederatedScope-LLM: A Comprehensive Package for Fine-tuning Large Language Models in Federated Learning (2024) · cites this paper as [26]
Experimental — accuracy not yet measured
41 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Extrinsic Object Hallucination
“CHAIR (Caption Hallucination Assessment with Image Relevance) metric, which calculates what proportion of object words generated are actually in the image according to the ground truth captions.”
§ Extrinsic Object Hallucination
“CHAIR 𝑖 measures per-instance object hallucination, i.e. what fraction of object instances in each generated caption are hallucinated.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 7.1 Hallucination Definition in Abstractive Summarization
“For instance, in Table 1, given the input article shown in the caption, an example of intrinsic hallucination is ' The Ebola vaccine was rejected by the FDA in 2019, '”
§ 2.1 Categorization
“in the abstractive summarization task from Table 1, the generated summary ' The first Ebola vaccine was approved in 2021 ' contradicts the source content”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Survey of Hallucination in Natural Language Generation
“The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions; (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, machine translation, and visual-language generation; and (3) hallucinations in large language models (LLMs)”
§ 1 INTRODUCTION
“The third part discusses this phenomenon in LLMs in Section 13.”
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 3×
“Recently, the rapid development of NLG has captured the imagination of many thanks to the advances in deep learning technologies, especially Transformer [189]-based models like BERT [29], BART [100], GPT-2 [149], and GPT-3 [16].”
BackgroundAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Sequence to Sequence Learning with Neural Networks
Sutskever, I., Vinyals, O., & Le, Q. V. (2014) · arXiv · doi:10.48550/arxiv.1409.3215 · Cited by 17,766
Advances in Neural Information Processing Systems 29
Teymur, O., Zygalakis, K., & Calderhead, B. (2016) · Cited by 13,887
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929 · cited 1×
“This work investigates the performances of GPT-3 [13], GPT-Neo/J [141], GPT-2 [105] and a T5-based model [106].”
BackgroundAdvances in Neural Information Processing Systems 14: Proceedings of the 2001 Conference
(2002) · The MIT Press eBooks · doi:10.7551/mitpress/1120.001.0001 · Cited by 8,890
Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.703 · Cited by 8,821 · cited 1×
“Recently, the development of NLG has conquered the imagination of many thanks to the advances in deep learning technologies, especially the Transformer [159]-based models like BERT [24], BART [83], GPT-2 [127], and GPT3 [13].”
BackgroundLewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2019) · arXiv · doi:10.48550/arxiv.1910.13461 · Cited by 8,816 · cited 1×
“Recently, the development of NLG has conquered the imagination of many thanks to the advances in deep learning technologies, especially the Transformer [159]-based models like BERT [24], BART [83], GPT-2 [127], and GPT3 [13].”
Background
Research works printed but not linked: 16
- [4] Joshua Albrecht and Rebecca Hwa. 2007. A Re-examination of Machine Learning Approaches for Sentence-Level MT Evaluation. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics . Association for Computational Linguistics, Prague, Czech Republic, 880-887. https://aclanthology.org/P07-1111 -
- [16] Andrzej Białecki, Robert Muir, Grant Ingersoll, and Lucid Imagination. 2012. Apache lucene 4. In SIGIR 2012 workshop on open source information retrieval . 17. -
- [53] Yilun Du, Shuang Li, Antonio Torralba, Joshua B Tenenbaum, and Igor Mordatch. 2023. Improving Factuality and Reasoning in Language Models through Multiagent Debate. In Forty-first International Conference on Machine Learning . -
- [121] Philipp Koehn and Rebecca Knowles. 2017. Six Challenges for Neural Machine Translation. In First Workshop on Neural Machine Translation . Association for Computational Linguistics, 28-39. -
- [145] Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation. In ICML . -
- [149] Tian Li, Ahmad Beirami, Maziar Sanjabi, and Virginia Smith. 2020. Tilted Empirical Risk Minimization. In International Conference on Learning Representations . -
- [164] Jiasen Lu, Jianwei Yang, Dhruv Batra, and Devi Parikh. 2018. Neural Baby Talk. In CVPR . -
- [165] Yukun Ma, Khanh Linh Nguyen, Frank Z Xing, and Erik Cambria. 2020. A Survey on Empathetic Dialogue Systems. Information Fusion 64 (2020), 50-70. -
- [175] Marianna J. Martindale, Marine Carpuat, Kevin Duh, and Paul McNamee. 2019. Identifying Fluently Inadequate Output in Neural and Statistical Machine Translation. In MTSummit . -
- [179] Kathleen McKeown. 1992. Text Generation . Cambridge University Press. -
- [209] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language Models are Unsupervised Multitask Learners. OpenAI Blog 1, 8 (2019), 9. -
- [238] Hui Su, Xiaoyu Shen, Sanqiang Zhao, Zhou Xiao, Pengwei Hu, Cheng Niu, and Jie Zhou. 2020. Diversifying Dialogue Generation with Non-Conversational Text. In 58th Annual Meeting of the Association for Computational Linguistics . ACL, 7087-7097. -
- [256] Zhaopeng Tu, Yang Liu, Lifeng Shang, Xiaohua Liu, and Hang Li. 2017. Neural Machine Translation with Reconstruction. In Thirty-First AAAI Conference on Artificial Intelligence . -
- [261] Neeraj Varshney, Wenlin Yao, Hongming Zhang, Jianshu Chen, and Dong Yu. 2023. A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generation. CoRR abs/2307.03987 (2023). https://doi.org/10.48550/arXiv.2307.03987 arXiv:2307.03987 - · arXiv · doi:10.48550/arxiv.2307.03987 · Paper page
- [285] Zeqiu Wu, Michel Galley, Chris Brockett, Yizhe Zhang, Xiang Gao, Chris Quirk, Rik Koncel-Kedziorski, Jianfeng Gao, Hannaneh Hajishirzi, Mari Ostendorf, et al. 2021. A Controllable Model of Grounded Response Generation. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 35. 14085-14093. -
- [291] Xinnuo Xu, Ondřej Dušek, Verena Rieser, and Ioannis Konstas. 2021. AGGGEN: Ordering and Aggregating while Generating. roceedings of the 59th Annual Meeting of the Association for Computational Linguistics (ACL2021) (2021). -
Cited by
Items 2,301–2,400 of 4,844
Xu, Y., Fan, S., & Jiang, H. (2026) · Decision Support Systems · doi:10.1016/j.dss.2025.114592 · Cited by 2
Self-citationSafeRAG: Secure Cloud-Based Retrieval-Augmented Generation for LLM-Empowered Voice Assistants
Chang, Y., Luan, T. H., Wang, S., & Wang, Y. (2026) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2026.3654070 · Cited by 2
“Neverthe-less, LLM-driven assistants still suffer from hallucination [7], [8], generating responses that are grammatically correct but factually inaccurate or unsupported, which significantly reduces their reliability.”
Information et intelligence artificielle
Dierickx, L., & Lindén, C. (2025) · De Boeck Supérieur eBooks · doi:10.3917/dbu.broud.2025.01.0031 · Cited by 1
Agentic AI Prompt Engineering: Advancing Generative AI Patterns Conceptually
Schneider, J. (2025) · International Conference on Information Technology and Digital Applications (ICITDA) · doi:10.1109/icitda68167.2025.11332344 · Cited by 1
“Additionally, hallucinations—instances where the model confidently produces plausible yet inaccurate or fabricated information—remain a pervasive challenge, compromising the reliability and factual correctness of outputs including reasoning traces [11], [12].”
Melillos, N., & Αγαπίου, Ά. (2026) · Heritage · doi:10.3390/heritage9010032 · Cited by 1
“Large language and vision–language models may generate plausible but incorrect descriptions, associations, or spatial inferences that are not grounded in the underlying data, potentially leading to misinterpretation of archaeological evidence [69,70].”
Active Personas for Synthetic User Feedback: A Design Science Study
Simaremare, M. E. S., & Edison, H. (2026) · Lecture notes in business information processing · doi:10.1007/978-3-032-14518-5_20 · Cited by 1
Large Language Model–Based Agents for Physical Activity and Cognitive Training: Scoping Review
Silacci, A., Giachetti, B., Angelini, L., Lopomo, N. F., Andreoni, G., Mugellini, E., Cherubini, M., & Caon, M. (2026) · JMIR AI · doi:10.2196/80123 · Cited by 2
“Once a user diverges from the scripted interaction, the role coherence can collapse, exposing the user to potential agent hallucinations [51].”
Chen, X., He, B., Lin, H., Han, X., Wang, T., Cao, B., Sun, L., & Sun, Y. (2026) · ACM Transactions on Information Systems · doi:10.1145/3788865 · Cited by 3
Erkan, M. H., Rahman, Ö. F., Güner, A., Ayyıldız, F., & Barbarus, E. (2026) · The European Research Journal · doi:10.18621/eurj.1839146 · Cited by 1
Retrieval Augmented Generation Enhanced RLHF Framework for Factuality Verification Improvement
Zhou, X. (2025) · IEEE International Conference on Frontiers Technology of Information and Computer (ICFTIC) · doi:10.1109/icftic68075.2025.11325091 · Cited by 1
“…into intrinsic (model-internal) and extrinsic (knowledge-gap) hallucinations and show that they become more severe under domain shift, temporal drift, or multi-hop reasoning conditions, directly affecting the reliability and traceability of outputs in question answering and summarization tasks [2].”
Securing Agentic AI: Threats, Risks, and Mitigation
Rashid, S. M. Z. U., Montasir, I., Haq, A., Ahmmed, M. T., & Alam, M. A. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-95-3701-3_44 · Cited by 3
Cai, W., Zhu, X., Sun, Z., Ye, A., Fu, G., & Zhang, R. (2026) · Advanced Engineering Informatics · doi:10.1016/j.aei.2026.104327 · Cited by 3
HADA: Human-AI Agent Decision Alignment Architecture
Pitkäranta, T., & Pitkäranta, L. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-15632-7_5 · Cited by 4
Medication Counseling with Large Language Models: Balancing Flexibility and Rigidity
Sabel, J., Wingren, M., Lundell, A., Andersson, S., Rosenberg, S., Hägglund, S., Estman, L. S., & Andtfolk, M. (2025) · IEEE International Conference on Agents (ICA) · doi:10.1109/ica67499.2025.00047 · Cited by 3
“However, this flexibility causes issues on the other end of the spectrum, as LLM systems have a tendency to hallucinate — that is, generate incorrect or unrealistic responses in relation to their source content [5].”
Xu, X., & Wu, J. (2026) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2026.3655508 · Cited by 10
“These interactions constitute the cornerstone of collaboration but simultaneously create pathways for error propagation [11].”
Wells, C. I., Han, A., Joseph, N., Varghese, C., O’Grady, G., & Bissett, I. P. (2026) · BMJ Quality & Safety · doi:10.1136/bmjqs-2025-019556 · Cited by 1
HFuzzer: Testing Large Language Models for Package Hallucinations via Phrase-based Fuzzing
Zhao, Y.-K., Wu, M.-H., Hu, X., & Xia, X. (2025) · International Conference on Automated Software Engineering · doi:10.1109/ase63991.2025.00225 · Cited by 8
“This issue has been extensively studied in natural language generation (NLG) [8].”
Uncovering Systematic Failures of LLMs in Verifying Code Against Natural Language Specifications
Jin, H., & Chen, H. (2025) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1109/ase63991.2025.00323 · Cited by 16
“This systematic failure may arise from inherent biases and hallucinations within LLMs [9], or from improper prompt designs.”
Alsayfi, M. S., Al-Qahtani, K. N., Aleisa, H. E. H., et al. (2026) · EDPACS · doi:10.1080/07366981.2026.2617746 · Cited by 1
Al-Khamees, H. A. A., Smadi, A. A., Alsmadi, M. K., ALKannad, A. A., Abugabah, A., Almusfar, L. A., & AlThani, B. (2026) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2026.200631 · Cited by 10
Understanding Artificial Intelligence (AI) for the Electrophysiologist
Ramanathan, C., & Trayanova, N. A. (2026) · Indian Pacing and Electrophysiology Journal · doi:10.1016/j.ipej.2026.01.010 · Cited by 1
“This limitation underlies the phenomenon of hallucination [34], in which generated outputs are syntactically coherent but factually incorrect or unsupported by evidence.”
Welfare Implications of Democratization in Content Creation: Generative AI and Beyond
Zou, T., Shi, Z., & Wu, Y. (2026) · Journal of Marketing Research · doi:10.1177/00222437261423540 · Cited by 15
“In the computer science literature, for example, Ji et al. (2023) offer a comprehensive overview of hallucinations in natural language generation.”
Revisiting computer authorship: a longitudinal perspective
Henrickson, L., & Leppänen, L. (2026) · AI & Society · doi:10.1007/s00146-025-02783-z · Cited by 4
“Computers ‘hallucinate’ while looking at images or writing text (Ji et al. 2023), and ‘dream’ while producing pictures (Mordvintsev, Olah and Tyka 2015).”
Chang, Y., Wang, S., Luan, T. H., Wang, Y., & Su, Z. (2025) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom66490.2025.00122 · Cited by 1
“However, despite their impressive capabilities, LLMs often suffer from the problem of hallucination [7], [8], where generated responses appear fluent but are factually incorrect or ungrounded.”
Ji, P., & Liu, X. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1725814 · Cited by 4
“Even in high-resource languages, LLMs still exhibit substantial biases in content generation (Ji et al., 2023), which may create cognitive confusion for older users who lack algorithmic literacy.”
Mallouhy, R. E., AlQahtani, H., & Lardhi, S. (2026) · IEEE Access · doi:10.1109/access.2026.3660770 · Cited by 3
“Despite these advancements, one of the primary concerns in using AI generated documentation in scientific fields is the risk of hallucinations, where AI models produce factually incorrect or misleading information [18].”
Ara, U., & Pandey, S. (2025) · Emerald Group Publishing Limited eBooks · doi:10.1108/s2040-724620250000022023 · Cited by 3
Integrating SysML V2 into a GRAG LLM Pipeline: Design, Implementation and Evaluation
Bader, E., Gross, J., Polanec, K., et al. (2025) · IEEE International Symposium on Systems Engineering · doi:10.1109/isse65546.2025.11369999 · Cited by 1
“As a result, their response often includes hallucinations, meaning they confidently write factually incorrect or fabricated information [4].”
Rizqullah, M., & Albassam, E. (2026) · IEEE Access · doi:10.1109/access.2026.3662817 · Cited by 5
“In addition, LLMs have been shown to exhibit hallucinations in which they assert that what is actually false is true [16], [17].”
Cheng, W., Yu, Z., & Qin, F. (2026) · Journal of Medical Systems · doi:10.1007/s10916-026-02349-6 · Cited by 1
“…to prospec - tively characterize systematic failure patterns—such as guideline-sensitive reasoning or terminology-dependent ambiguities—thereby supporting the development of more representative safety benchmarks and adversarial test sets, as suggested in prior medical AI safety literature [4].”
Song, Z., Kong, X., Bao, X., Zhou, Y., Jiao, J., Liu, S., Zhou, Y., & Qi, H. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.131627 · Cited by 8
“This issue poses a significant barrier in knowledge-intensive applications, such as medical diagnosis, where factual accuracy is paramount (Ji et al., 2023; Huang et al., 2025; Luo et al., 2026; Płonka et al., 2025; Chi et al., 2026; L. Li et al., 2025).”
Wang, Z., Li, L., Chen, L., Xu, X., Li, M., & Su, Y. (2025) · International Conference on Wireless Communication, Networking and Internet of Things (WCNIoT) · doi:10.1109/wcniot67424.2025.11381315 · Cited by 1
“However, along with the rapid adoption of generative AI, an inherent and widely discussed challenge—hallucination [16], [17]—has drawn increasing attention.”
Li, Y., Lv, M., Cheng, L., Xie, Y., Gu, Q., He, H., Chen, Z., Fang, D., & Zhou, X. (2026) · European Journal of Radiology · doi:10.1016/j.ejrad.2026.112734 · Cited by 2
Günay, E. E., Chern, W., Elhabashy, A. E., Kremer, P. A., Haapala, K. R., Kim, K.-Y. J., & Kremer, G. E. O. (2026) · Journal of Computing and Information Science in Engineering · doi:10.1115/1.4071183 · Cited by 3
Calibrated Trust in Dealing with LLM Hallucinations: A Qualitative Study
Ryser, A., Allwein, F., & Schlippe, T. (2025) · International Conference on Foundation and Large Language Models (FLLM) · doi:10.1109/fllm67465.2025.11391250 · Cited by 7
“Hallucinations are defined as content generated by LLMs that is linguistically convincing but factually incorrect [1], [12].”
Language Modeling and Large Language Models
Yu, J. S., & Yao, Y. (2026) · Springer eBooks · doi:10.1007/978-981-95-4632-9_2 · Cited by 2
TermiTeach: Natural Language Command Line Interface Tutoring Via Generative AI Simulations
Singh, V. K. (2025) · International Symposium on Embedded Computing and System Design · doi:10.1109/ised67359.2025.11405208 · Cited by 1
“However, this approach also introduces challenges inherent to the use of LLMs. • Hallucination: The risk of the LLM inventing command outputs or file system states is a concern [13].”
Krainovskikh, V., & Samigulin, T. (2026) · IEEE Access · doi:10.1109/access.2026.3668910 · Cited by 3
“Relative to DFT, RAG enhances factual accuracy, mitigates hallucination risk via retrieval grounding, rather than eliminating it entirely, and enables transparent source attribution [12], [13] — capabilities that are particularly useful for technical and safety-sensitive applications [10].”
Review of generative AI for lesion localization and automatic report generation
Sun, K., Wang, Y., Huang, S., Liu, Y., Chen, H., Lian, Z., Tan, M., Lin, X., Zhou, W., Qiu, Z., et al. (2026) · Med-X · doi:10.1007/s44258-026-00076-8 · Cited by 2
Liu, D., & Yuan, Y. (2026) · Asian Journal of Communication · doi:10.1080/01292986.2026.2639664 · Cited by 1
“On the contrary, it may produce information that appears accurate but is factually incorrect or misleading (i.e. AI hallucinations; Ji et al., 2023), leading to health risks for users who overly rely on AI-generated health information without carefully evaluating it (Jin et al., 2025; Shi et al.,…”
Push and Pull : Defending against Retrieval Poisoning Attacks via Embedding Space Reshaping
He, L., Zhang, X., Liu, Q., Li, C., Liu, Z., Zhou, P., & Su, S. (2026) · ACM Transactions on Information Systems · doi:10.1145/3800948 · Cited by 6
Augmenting RAG with nonnegative matrix factorization-driven semantic chunking in embedding space
Abdolinejad, F., & Eftekhari, M. (2026) · The Journal of Supercomputing · doi:10.1007/s11227-026-08370-3 · Cited by 1
DSRAG: A Domain-Specific Retrieval Framework Based on Document-Derived Multimodal Knowledge Graph
Yang, M., Ren, Y., Opoku, D. O., Li, R., Ren, P., & Xing, C. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-6426-2_46 · Cited by 4
Zhang, L., Huang, D., Yu, W., Yang, W., Feng, D., Xie, H., & Zhang, H. (2026) · Journal of Pipeline Science and Engineering · doi:10.1016/j.jpse.2026.100473 · Cited by 1
Ram, A. M., Saravi, S., Yang, Y., Batmaz, F., Derakhshan, P., Nevisi, H., & Storey, G. (2026) · Education and Information Technologies · doi:10.1007/s10639-026-13934-0 · Cited by 9
“AI is used only as a sparring partner (Stokkink, 2025 assessments are intended to measure and can encourage over-reliance on automation (Ji et al., 2023; Kasneci et al., 2023; Risko & Gilbert, 2016).”
InfluentialZhang, X. (2026) · AI & Future Society · doi:10.63802/afs.v2.i1.208 · Cited by 1
CausalAgent: A Conversational Multi-Agent System for End-to-End Causal Inference
Zhu, J., Chen, W., & Cai, R. (2026) · ACM International Conference on Intelligent User Interfaces (IUI) · doi:10.1145/3742414.3794777 · Cited by 1
PubMedRAG: Retrieval-Augmented Generation for Biomedical Question Answering
Gupta, H., Sivakumar, A., Fu, H., & Qiang, Y. (2025) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw69685.2025.00352 · Cited by 3
“First, the training data used for these models quickly become outdated, limiting their ability to incorporate the latest medical research and clinical guidelines needed for accurate responses in healthcare settings [6].”
Hybrid retrieval generation for structured reasoning with large language models
Muthusami, R., & Saritha, K. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01059-9 · Cited by 4
Caridi, A., Giovannini, M., & Celsi, L. R. (2026) · Electronics · doi:10.3390/electronics15061155 · Cited by 7
“The objective is not to assume that agreement implies correctness but to use model diversity as a practical mechanism for detecting unstable outputs, unsupported values, and source inconsistencies before these propagate to KPI computation and valuation [29].”
Wu, J., Chen, S., Chen, P., Du, W., Gao, H. Z., Wang, L., & Hansen, P. (2026) · Design and Artificial Intelligence · doi:10.1016/j.daai.2026.100063 · Cited by 4
Öztaş, B., & Kösesoy, İ. (2026) · Clinica Chimica Acta · doi:10.1016/j.cca.2026.120974 · Cited by 2
Pearson, S., Reyburn, M., Foley, C., Finch, A., Bench, S. D., Bonnici, T., & Rose, L. (2026) · JMIR Formative Research · doi:10.2196/90139 · Cited by 1
“However, LLMs may hallucinate, cite unverifiable sources, and produce materials misaligned with literacy needs [4].”
Si, N., Shen, Q., Wu, S., et al. (2025) · International Conference on Computers and Artificial Intelligence Technology (CAIT) · doi:10.1109/cait68620.2025.11424748 · Cited by 1
“(2) A common understanding exists that LLMs frequently encounter hallucinations in their answer generation, [15], and excessive creativity is deemed inappropriate in urban research.”
Khanayev, A., Balcı, M., Phutkaradze, I., & Khudaverdiyeva, A. (2026) · New Perspectives · doi:10.1177/2336825x261436400 · Cited by 1
Leite, M. V., Abe, J. M., Nääs, I. D. A., & Souza, M. L. H. (2026) · AgriEngineering · doi:10.3390/agriengineering8030114 · Cited by 6
DRGO:dynamic retrieval and generation optimization based on fine-grained confidence evaluation
Wang, Z., Yang, T., Wang, X., Zhang, Y., Qiu, H., & Feng, J. (2026) · Journal of Intelligent Information Systems · doi:10.1007/s10844-026-01036-9 · Cited by 1
“However, despite the robust language processing capabilities exhibited by LLMs, the content they generate may still exhibit "hallucination" phenomena, wherein the model produces information that appears plausible but is in fact erroneous or fabricated (Ji et al., 2023; Cao et al., 2020).”
One Line at a Time: Scaffolding Reflective Code Evaluation through Structured AI Assistance
Qiu, V., Parker, L., Riegel, K., Giacaman, N., Denny, P. C., MacNeil, S., & Prather, J. E. (2026) · Australasian Computing Education Conference · doi:10.1145/3786228.3786243 · Cited by 4
Yin, C., Sun, X., Dai, A., Ye, X., Lu, Y., Wang, W., Chen, Y., Jiang, H., Yu, J., Chong, S., et al. (2026) · APL Bioengineering · doi:10.1063/5.0299327 · Cited by 2
Sidhu, R., Selvamogan, A., & Boddy, A. (2026) · Annals of The Royal College of Surgeons of England · doi:10.1308/rcsann.2026.0021 · Cited by 3
McBride, R. (2026) · Strategy Science · doi:10.1287/stsc.2025.0439 · Cited by 6
Large language models for car following in automated driving: Opportunities and challenges
Pang, S., Patel, D., Liu, Y., Wang, S., Lin, C., He, S., & Li, T. (2026) · Artificial Intelligence for Transportation · doi:10.1016/j.ait.2026.100047 · Cited by 4
Learned Hallucination Detection in Black-Box LLMs Using Token-Level Entropy Production Rate
Moslonka, C., Randrianarivo, H., Garnier, A., & Malherbe, E. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21289-4_8 · Cited by 8
When Semantic Similarity Fails: Analyzing Metric Divergence in Legal RAG Evaluation
SILVA, E. D. D. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21324-2_43 · Cited by 1
Bharucha, D., Palanisamy, N., & Reddy, M. (2026) · Biennial International Conference on Nascent Technologies in Engineering (ICNTE) · doi:10.1109/icnte66387.2026.11437450 · Cited by 2
“The only problem with their adoption in compliance of high stakes is the likelihood of the models producing incorrect statements or "hallucination" [13].”
AgriIR: A Scalable Framework for Domain-Specific Knowledge Retrieval
Seal, S. B., Poddar, A., Mishra, A., & Roy, D. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21324-2_37 · Cited by 2
“Recent work emphasizes citation reliability [17], with studies showing GPT-3.”
From Hallucination to Certainty: Meta-Knowledge Guided Self-Correcting Large Language Models
Zhang, W., Dai, G., Luo, D., Wang, Y., & YE, C. (2026) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3797906 · Cited by 5
SpecFSM: Extracting and Repairing Finite State Machines from Protocol Specification Documents
Li, X., Huang, D., Guan, J., & Li, H. (2026) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1145/3832783.3834331
When AI evaluates its own work: Validating learner-initiated, AI-generated physics practice problems
Geisler, T., & Kortemeyer, G. (2026) · Physical Review Physics Education Research · doi:10.1103/9hh9-vt4d · Cited by 4
“…(“hallucinations” in this context refers to instances where an LLM generates plausible-sounding but inaccurate or entirely fabricated information [68]). realistic-data-used : There was only one problem that had unrealistic data, namely a problem with a freight train that has a total mass of…”
Klein, A., Song, J., Chim, J., Keren, L., Triantafyllopoulos, A., Schuller, B. W., Liakata, M., & Atzil–Slonim, D. (2026) · JMIR Formative Research · doi:10.2196/71230 · Cited by 1
Epistemic Limits of Hallucination Mitigation in Large Language Models
Zhang, Y. (2026) · Foundations of Science · doi:10.1007/s10699-026-10030-x · Cited by 4
Artificial Intelligence-Based Digital Transformation
Volpentesta, T. (2026) · Emerald Group Publishing Limited eBooks · doi:10.1108/978-1-80592-096-020261002 · Cited by 1
Singh, R., Venugopal, V. K., Daga, R., et al. (2026) · European Journal of Radiology Artificial Intelligence · doi:10.1016/j.ejrai.2026.100088 · Cited by 1
Chen, P., Huang, Y., Wu, T., & Lee, H. (2025) · International Conference on Modern Educational Technology (ICMET) · doi:10.1109/icmet67594.2025.11451842 · Cited by 4
L-PRISMA: An Extension of PRISMA in the Era of Generative Artificial Intelligence (GenAI)
Shailendra, S., Kadel, R., Sharma, A., Islam, M. T., & Saxena, U. R. (2025) · International Conference on Modern Educational Technology (ICMET) · doi:10.1109/icmet67594.2025.11452106 · Cited by 2
“It must also be noted that since LLMs have the tendency to hallucinate [22], [23] and generate false results confidently, caution needs to be exercised to review all outcomes by a human during the process.”
Pedraja-Rejas, L., Acosta, K., Rodríguez-Ponce, E., & Muñoz-Fritis, C. P. (2026) · Systems · doi:10.3390/systems14040403 · Cited by 2
KOSMOS: Ontology-Based Knowledge Graph Scaffolding for Medical Documentation Generation
Henry, R., & Gong, J. (2026) · Information · doi:10.3390/info17040355 · Cited by 2
Huh, Y., Noh, M., Ji, D., Oh, Y., & Sun, S. (2026) · Sensors · doi:10.3390/s26082316 · Cited by 2
Ren, Y., Osborn, J. D., Brost, E., Yu, H., Abidi, S. H., Go, Y., Li, S., Park, D. K., Huang, Y., Luo, X., et al. (2026) · International Journal of Modern Physics A · doi:10.1142/s0217751x26500867 · Cited by 3
Zhang, G., Sun, G., Xia, M., & Liang, R. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790824 · Cited by 5
Sustainable Materials Design With Multi‐Modal Artificial Intelligence
Xu, T., Wei, T., Ge, Y., Peng, B., Li, Y., Wang, M., Wen, P., Yang, C., & Wei, Y. (2026) · Advanced Science · doi:10.1002/advs.202524273 · Cited by 7
“Documented failure modes include hallucinated facts, error amplification from biased or incomplete corpora, and missing or violated physical constraints [ 170, 171 ].”
Xu, T., & Suzuki, H. (2026) · Cogent Business & Management · doi:10.1080/23311975.2026.2653306 · Cited by 2
Medicherla, C., Kulkarni, C., Patel, T., Shivam, S., Nutalapati, V., Parikh, M., Kumari, M., & Elengovan, A. K. (2026) · International Seminar on Intelligent Business and Edge-Computing Research (ISIBER) · doi:10.1109/isiber68248.2026.11470068 · Cited by 4
“Following Ji et al. [11], we identify four hallucination categories specific to RAG systems: (1) Factual hallucinations that directly contradict information in context (e.g., stating Sydney is Australia's capital when context says Canberra), (2) Extractive hallucinations that correctly extract…”
Correcting AI for innovation: how networks shape knowledge correction capability in hybrid systems
Huang, Y. (2026) · Asia Pacific Journal of Marketing and Logistics · doi:10.1108/apjml-08-2025-1557 · Cited by 2
RAMASC: A retrieval-augmented multi-agent framework for automated structural calculation
Choi, K., Jeong, M., Kim, T., Kim, S., Baek, S., & Kim, H. (2026) · Advanced Engineering Informatics · doi:10.1016/j.aei.2026.104698 · Cited by 5
A Graph-Enhanced Defense Framework for Explainable Fake News Detection with LLM
Wang, B., Ma, J., Lin, H., Yang, Z., Yang, R., Tian, Y., & Chang, Y. (2026) · ACM Transactions on Information Systems · doi:10.1145/3810244 · Cited by 4
Wang, Y., & Ge, Y. (2026) · Information Systems Research · doi:10.1287/isre.2024.1518 · Cited by 3
Semantic Reformulation Entropy for Robust Hallucination Detection in QA Tasks
Tong, C., Zhang, Q., Jiang, L., Liu, Y., Sun, N., & Li, W. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11460452 · Cited by 5
Li, J., Chang, Y., Liu, Y., Wang, Y., Ou, B., Lin, S., Tang, S.-K., & Zheng, L. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1804478 · Cited by 5
“…general-purpose LLMs have some eectiveness in the medical field, they still fall far short of expert levels (5), and sometimes they may produce incorrect or misleading content due to “hallucinations,” which may be related to the quality of training data, model architecture, training processes (6).”
Conversational recommender systems with large language models: A review of current developments
Kim, Y.-H., & Cho, Y.-S. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.132508 · Cited by 6
The Development of Information Retrieval Model for Hadith Data from Knowledge Graph
Norhisham, N. N., Muhammad, N., Ismail, M. I. F., & Fabregas, A. C. (2026) · International Conference on Electronics, Computer and Computation · doi:10.1109/icecco67619.2026.11488814 · Cited by 1
“However, these models are known to hallucinate, i.e., generate factually incorrect responses when they lack structured, domain-specific knowledge (Ji et al., 2023; Dziri et al., 2022).”
Liu, Z., Ding, Y., Chen, J., Yan, Z., Cheng, X., Zhou, W., Fu, P., & Wang, Z. (2026) · JMIR Medical Informatics · doi:10.2196/97520 · Cited by 1
“Second, the hallucination problem intrinsic to LLMs poses particular dangers in clinical settings [10].”
Understanding Hallucinations in Large Visual and Language Models
Ho, Z. Y., Liang, S., & Tao, D. (2026) · ACM Computing Surveys · doi:10.1145/3811409 · Cited by 11
Wang, R., Li, Y., Feng, X., & Feng, X. (2026) · International Journal of Clinical Pharmacy · doi:10.1007/s11096-026-02138-8 · Cited by 1
Xiong, J., & Li, F. (2026) · Educational Research Review · doi:10.1016/j.edurev.2026.100789 · Cited by 6
Hallucination detection in large language models based on multi-granularity consistency
Zhang, J., Ma, X., Yang, Q., Cui, T., Chen, X., Lin, Q., & Zhao, B. (2026) · Journal of Intelligent Information Systems · doi:10.1007/s10844-026-01047-6 · Cited by 1
“In NLP research, hallucination refers to the phenomenon that the textual content generated by a model is inconsistent with the source information, lacks factual basis, or is difficult to validate (Ji et al., 2023).”
CHEN, C. X. R., Xie, S., Luo, Z., Hu, Z., Du, X., Huang, D., Hu, H., & Peng, S. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1817855 · Cited by 2
“This phenomenon of “hallucinations,” generating content that contradicts facts or context, is an intrinsic property of the model architecture, not a mere engineering oversight (55).”
Yan, G., Zhu, C., Li, T., Yip, J., & Ni, J. (2026) · PLoS ONE · doi:10.1371/journal.pone.0347653 · Cited by 1
“…improprieties, insufficient methodological traceability, and diluted originality—especially when usage is not accompanied by labeling, process logs, and verification procedures [12]; under such conditions, the boundary between tool output and learners’ own academic competence can blur [13].”
Taioli, F., Zorzi, E., Franchi, G., Castellini, A., Farinelli, A., Cristani, M., & Wang, Y. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01745 · Cited by 25
“Hallucinations, biases, reasoning failures and the generation of unfaithful text by LLMs are well-known issues [18].”
Ho, T.-L., & Lam, T.-P. (2026) · Applied Sciences · doi:10.3390/app16094400 · Cited by 5
Metadata
- Authors
- Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P.
- Affiliations
- Hong Kong University of Science and Technology (HK)
- ORCID
- Venue
- ACM Computing Surveys
- Published
- 2023-03-03
- Volume
- 55
- Issue
- 12
- Pages
- 1-38
- ISSN
- 0360-0300 (ISSN-L)
- 1557-7341
- OpenAlex
- W4309674289
- Semantic Scholar
- 3def68bd0f856886d34272840a7f81588f2bc082
- Cited by
- 5,025
- References
- 111
- Citations by year
- 2026: 2296 · 2025: 1441 · 2024: 817 · 2023: 453
- FWCI
- 224.36
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- ACM Computing Surveys (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
- Keywords
- hallucinations, natural language generation, hallucination mitigation, abstractive summarization, dialogue generation, generative question answering, data-to-text generation, machine translation, large language models, transformer models
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.




