大模型智能体幻觉难题：成因、风险与应对

徐琦 孙智蒲

doi:10.19483/j.cnki.11-4653/n.2025.05.001

大模型智能体幻觉难题：成因、风险与应对

doi: 10.19483/j.cnki.11-4653/n.2025.05.001

徐琦孙智蒲

中国传媒大学媒体融合与传播国家重点实验室新媒体研究院

基金项目:

本文相关研究由“国家广播电视总局部级社科研究项目资助”（项目名称：《AIGC 大模型在广电行业内容生产和传播中的应用研究，项目编号：GD2415）。

详细信息

作者简介:
徐琦孙智蒲：徐琦（1982—），女，中国传媒大学媒体融合与传播国家重点实验室新媒体研究院副研究员、硕士研究生导师，研究方向为智能传播、媒体融合、数字人文与新媒体等；孙智蒲（2001—），男，硕士研究生，研究方向为智能媒体、人机传播、媒体融合。

计量
- 文章访问数: 10
- HTML全文浏览量: 0
- PDF下载量: 0
- 被引次数: 0
出版历程

摘要

摘要: 【目的】大模型智能体幻觉及风险问题日益凸显，深入解析其成因、风险表现及其应对措施具有重要的理论与应用意义。【方法】面向新闻传播领域理论与应用需求，本研究主要基于对跨学科文献研究与理论辨析开展。【结果】智能体幻觉意指模型层在生成上难以避免地出现了生成内容不合逻辑或不忠于所提供的源内容等一系列错误，主要分为事实性幻觉与忠实性幻觉两类。前者包括事实错误、编造和忽视，后者涵盖意图、上下文和逻辑不一致。在下游应用中，幻觉风险广泛存在于机器翻译、问答系统、对话、摘要、知识图谱和视觉问答等任务，表现为翻译偏离、不完整回答、信息扭曲等，危及内容真实性和准确性。【结论】为应对幻觉难题，传媒业首先要从认知层面来强化风险意识与技术素养，技术上可采用检索增强生成和事实性解码策略，流程上要完善人机协同流程，增强校验与多维评估体系，以平衡智能体效能与可靠性。
- 智能体 /
- 大模型 /
- 幻觉 /
- 智能媒体 /
- 智能传播

HTML全文

参考文献(56)

[1]	决明子 . DeepSeek 正在中文互联网建造“幻觉长城”[EB/OL]. （2025-02-07）[2025-04-25]. https：//mp.weixin.qq.com/s/aMy99RcCq62D9JvTgTUi7A.
[2]	Kalai A T，Vempala S S. Calibrated language models must hallucinate[C]. Proceedings of the 56th Annual ACM Symposium on Theory of Computing，2024：160-171.
[3]	Vectara. DeepSeek-R1 hallucinates more than DeepSeekV3[EB/OL]. （2025-01-30） [2025-04-25]，https：//www.vectara.com/blog/deepseek-r1-hallucinates-morethan-deepseek-v3.
[4]	-42.
[4]	Nicola J. AI hallucinations can’t be stopped—but these techniques can limit their damage[J].Nature. 2025，637（8047）：778-780.
[5]	张铮，刘晨旭 . 大模型幻觉：人机传播中的认知风险与共治可能 [J]. 苏州大学学报（哲学社会科学版），2024，45 （5）：171-180.
[6]	经羽伦，张殿元 . 生成式 AI 幻象的制造逻辑及其超真实建构的文化后果 [J]. 山东师范大学学报（社会科学版），2024，69（5）：113-126.
[7]	张新生，王润周，马玉龙 . AIGC 背景下虚假信息治理挑战、机会与策略研究 [J/OL]. 情报科学，1-23[2025-06-05].http：//kns.cnki.net/kcms/detail/22.1264.G2.20241111.1002.024.html.
[8]	Chakraborty N，Ornik M，Driggs-Campbell K. Hallucination detection in foundation models for decisionmaking： A flexible definition and review of the state of the art[J]. ACM Computing Surveys，2025，52（7）：1-35.
[9]	Wu J，Gan W，Chen Z，et al. Multimodal large language models：A survey[C]. 2023 IEEE International Conference on Big Data. IEEE，2023：2247-2256.
[10]	Xi Z，Chen W，Guo X，et al. The rise and potential of large language model based agents： A survey[J]. Science China Information Sciences，2025，68（2）：101-121.
[11]	Gong R，Huang Q，Ma X，et al. MindAgent：Emergent Gaming Interaction[C]. Findings of the Association for Computational Linguistics：NAACL 2024，2024：3154-3183.
[12]	Zhang J，Huang J，Jin S，Lu S. Vision-language models for vision tasks：A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence，2024，46（8）：5625-5644.
[13]	El-Mallakh R S，Walker K L. Hallucinations，psuedohallucinations，and parahallucinations[J]. Psychiatry：Interpersonal and Biological Processes，2010，73（1）：
[14]	Chakraborty N，Ornik M，Driggs-Campbell K. Hallucination detection in foundation models for decisionmaking： A flexible definition and review of the state of the art[J]. ACM Computing Surveys，2025，52（7）：1-35.
[15]	Sahoo P，Meharia P，Ghosh A，et al. A ComprehensiveSurvey of Hallucination in Large Language，Image，Video and Audio Foundation Models[C]. Findings of the Association for Computational Linguistics：EMNLP 2024. 2024：11709-11724.
[16]	Chen X，Wang C，Xue Y，et al. Unified Hallucination Detection for Multimodal Large Language Models[C]. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics. 2024，1: 3235-3252.
[17]	Hicks M T，Humphries J，Slater J. ChatGPT is bullshit[J]. Ethics and Information Technology，2024，26（2）：1-10.
[18]	Huang L，Yu W，Ma W，et al. A survey on hallucination in large language models：Principles，taxonomy，challenges，and open questions[J]. ACM Transactions on Information Systems，2025，43（2）：1-55.
[19]	Ji Z，Lee N，Frieske R，et al. Survey of hallucination in natural language generation[J]. ACM computing surveys，2023，55（12）：1-38.
[20]	Huang L，Yu W，Ma W，et al. A survey on hallucination in large language models：Principles，taxonomy，challenges，and open questions[J]. ACM Transactions on Information Systems，2025，43（2）：1-55.
[21]	Chen X，Wang C，Xue Y，et al. Unified Hallucination Detection for Multimodal Large Language Models[C]. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics. 2024，1: 3235-3252.
[22]	Guerreiro N M，Alves D M，Waldendorf J，et al. Hallucinations in large multilingual translation models[J]. Transactions of the Association for Computational Linguistics，2023，11：1500-1517.
[23]	Zheng L，Chiang W L， Sheng Y，et al. Judging llmas-a-judge with mt-bench and chatbot arena[J]. Advances in Neural Information Processing Systems，2023，36：46595-46623.
[24]	Adlakha V，Ghader B P，Lu X H，et al. Evaluating correctness and faithfulness of instruction-following models for question answering[J]. Transactions of the Association for Computational Linguistics 2024，12：681-699.
[25]	Dziri N，Milton S，Yu M，et al. On the Origin of Hallucinations in Conversational Models：Is it the Datasets or the Models?[C]. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics：Human Language Technologies， 2022：5271-5285.
[26]	Das S，Saha S，Srihari R K. Diving Deep into Modes of Fact Hallucinations in Dialogue Systems[C]. Findings of the Association for Computational Linguistics: EMNLP 2022，2022：684-699.
[27]	Qiu Y，Ziser Y，Korhonen A，et al. Detecting and Mitigating Hallucinations in Multilingual Summarisation[C]. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing，2023：8914-8932.
[28]	Yuan S，Faerber M. Evaluating Generative Models for Graph-to-Text Generation[C]. Proceedings of the 14th International Conference on Recent Advances in Natural Language Processing，2023：1256-1264.
[29]	Li Y， Du Y， Zhou K， et al. Evaluating Object Hallucination in Large Vision-Language Models[C]. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing，2023：292-305.
[30]	刘泽垣，王鹏江，宋晓斌，等 . 大语言模型的幻觉问题研究综述 [J]. 软件学报，2025，36（3）：1152-1185.
[31]	Lebret R，Grangier D，Auli M. Neural Text Generation from Structured Data with Application to the Biography Domain[C]. Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing，2016：1203-1213.
[32]	Lee K，Ippolito D，Nystrom A，et al. Deduplicating Training Data Makes Language Models Better[C]. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics，2022，1：8424-8445.
[33]	Rashkin H，Reitter D，Tomar G S，et al. Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features[C]. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing，2021，1：704-718.
[34]	Das B C，Amini M H，Wu Y. Security and privacy challenges of large language models：A survey[J]. ACM Computing Surveys，2025，57（6）：1-39.
[35]	Lin S，Hilton J，Evans O. TruthfulQA：Measuring How Models Mimic Human Falsehoods[C]. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics，2022，1：3214-3252.
[36]	Kasai J，Sakaguchi K，Le Bras R，et al. Realtime qa：What’s the answer right now?[J]. Advances in neuranformation processing systems，2023，36：49025-49043.
[37]	Paullada A，Raji I D，Bender E M，et al. Data and its （dis） contents：A survey of dataset development and use in machine learning research[J]. Patterns， 2021， 2（11）：1-14.
[38]	Gekhman Z，Yona G，Aharoni R，et al. Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations?[C]. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing，2024: 7765-7784.
[39]	Bhattacharya P，Prasad V K，Verma A，et al. Demystifying ChatGPT：An in-depth survey of OpenAI’s robust large language models[J]. Archives of Computational Methods in Engineering，2024：1-44.
[40]	Wang C，Sennrich R. On Exposure Bias， Hallucination and Domain Shift in Neural Machine Translation[C]. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020：3544-3552.
[41]	Zhang M，Press O，Merrill W，et al. How Language Model Hallucinations Can Snowball[C]. International Conference on Machine Learning，2024: 59670-59684.
[42]	Yang Y，Chern E，Qiu X，et al. Alignment for honesty[J]. Advances in Neural Information Processing Systems，2024，37：63565-63598.
[43]	Cotra， Ajeya. Why AI alignment could be hard with modern deep learning [EB/OL]. （2025-09-21）[2025-04-25]. Cold Takes. https://www.cold-takes.com/whyai-alignment-could-be-hard-with-modern-deeplearning/.
[44]	Fan A，Lewis M，Dauphin Y. Hierarchical Neural Story Generation[C]. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics，2018，1: 889-898.
[45]	Alves D，Guerreiro N，Alves J，et al. Steering Large Language Models for Machine Translation with Finetuning and In-Context Learning[C]. Findings of the Association for Computational Linguistics：EMNLP 2023，2023: 11127-11148.
[46]	Yang Z，Dai Z，Salakhutdinov R，et al. Breaking the Softmax Bottleneck：A High-Rank RNN LanguageModel[C]. International Conference on Learning Representations，2018：1-18.
[47]	Yuan Y，Wang W，Guo Q，et al. Does chatgpt know that it does not know? evaluating the black-box calibration of chatgpt[C]. Proceedings of the 2024 Joint International Conference on Computational Linguistics， Language Resources and Evaluation （LREC-COLING 2024），2024：5191-5201.
[48]	Tihanyi N， Bisztray T， Ferrag M A， et al. How secure is AI-generated code: a large-scale comparison of large language models[J]. Empirical Software Engineering， 2025， 30（2）: 1-42.
[49]	全会 . 冲击·融合·协同：ChatGPT 对传媒业的影响刍议 [J]. 中国广播电视学刊，2023，（09）：17-21.
[50]	Nicola J. AI hallucinations can’t be stopped—but these techniques can limit their damage[J].Nature. 2025，637（8047）：778-780.
[51]	Katzenbach C，Pentzold C，Otero P V. Smoothing out smart tech’s rough edges: Imperfect automation and the human fix[J]. Human-Machine Communication，2024，7：23-44.
[52]	郭全中，苏刘润薇，彭子滔 . 2023—2024 年传媒业大模型应用报告 [J]. 中国传媒科技，2025，（1）：6-10.
[53]	李子甜 . 工具性收益与系统性风险：新闻从业者的人工智能新闻技术认知 [J]. 新闻大学，2022（11）：29-42+117.
[54]	Lee M. A mathematical investigation of hallucination and creativity in GPT models[J]. Mathematics，2023，11（10）：2320.
[55]	Huang L，Yu W，Ma W，et al. A survey on hallucination in large language models：Principles，taxonomy，challenges，and open questions[J]. ACM Transactions on Information Systems，2025，43（2）：1-55.