Large Model-Based Agents: State-of-the-Art, Cooperation Paradigms, Security and Privacy, and Future Trends
Open accessThe DOI matches its Crossref record; 66 of the 66 references it cites were found in a registry or an index.
Abstract
With the rapid advancement of large models (LMs), the development of general-purpose intelligent agents powered by LMs has become a reality. It is foreseeable that in the near future, LM-driven general AI agents will serve as essential tools in production tasks, capable of autonomous communication and collaboration without human intervention. This paper investigates scenarios involving the autonomous collaboration of future LM agents. We review the current state of LM agents, the key technologies enabling LM agent collaboration, and the security and privacy challenges they face during cooperative operations. To this end, we first explore the foundational principles of LM agents, including their general architecture, key components, enabling technologies, and modern applications. We then discuss practical collaboration paradigms from data, computation, and knowledge perspectives to achieve connected intelligence among LM agents. After that, we analyze the security vulnerabilities and privacy risks associated with LM agents, particularly in multi-agent settings, examining underlying mechanisms and reviewing current and potential countermeasures. Lastly, we propose future research directions for building robust and secure LM agent ecosystems.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Mastering the game of Go without human knowledge
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., et al. (2017) · Nature · doi:10.1038/nature24270 · Cited by 9,084
Survey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025 · cited 1×
“For LLMs, source-reference divergence occurs due to heuristic data collection, implying that the selected sources may lack sufficient or accurate information for content generation [191].”
Huang, L., Yu, W., Ma, W., Zhong, W., Feng, Z., Wang, H., Chen, Q., Peng, W., Feng, X., Qin, B., et al. (2025) · ACM Transactions on Information Systems · doi:10.1145/3703155 · Cited by 4,688 · cited 3×
“LM agents are susceptible to “hallucinations”, where their LMs generate outputs that are plausible but incorrect and not grounded in reality, potentially spreading misinformation in multi-agent environments [28].”
Generative Agents: Interactive Simulacra of Human Behavior
Park, J. S., O'Brien, J., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023) · Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology · doi:10.1145/3586183.3606763 · Cited by 2,187 · cited 1×
“Role-playing [60], [98] is a typical method for facilitating cooperation among LM agents.”
Generative Agents: Interactive Simulacra of Human Behavior
Park, J. S., O'Brien, J. C., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023) · arXiv · doi:10.48550/arxiv.2304.03442 · Cited by 2,173
A survey on large language model based autonomous agents
Wang, L., Ma, C., Feng, X., Zhang, Z., Yang, H., Zhang, J., Chen, Z., Tang, J., Chen, X., Lin, Y., et al. (2024) · Frontiers of Computer Science · doi:10.1007/s11704-024-40231-1 · Cited by 1,941
A Survey on Large Language Model based Autonomous Agents
Wang, L., Ma, C., Feng, X., Zhang, Z., Yang, H., Zhang, J., Chen, Z., Tang, J., Chen, X., Lin, Y., et al. (2023) · arXiv · doi:10.48550/arxiv.2308.11432 · Cited by 1,925
Large Language Model based Multi-Agents: A Survey of Progress and Challenges
Guo, T., Chen, X., Wang, Y., Chang, R., Pei, S., Chawla, N. V., Wiest, O., & Zhang, X. (2024) · arXiv · doi:10.48550/arxiv.2402.01680 · Cited by 1,713
Self-citationA survey on large language model (LLM) security and privacy: The Good, The Bad, and The Ugly
Yao, Y., Duan, J., Xu, K., Cai, Y., Sun, Z., & Zhang, Y. (2024) · High-Confidence Computing · doi:10.1016/j.hcc.2024.100211 · Cited by 1,083 · cited 1×
“Forthe security and privacy of LLMs and agents, Yao et al. [40] provide a comprehensive review of LLMs’ dual role in cybersecurity, examining their potential to enhance security and privacy (“the good”), the risks and threats they introduce (“the bad”), and their inherent vulnerabilities (“the…”
The Rise and Potential of Large Language Model Based Agents: A Survey
Xi, Z., Chen, W., Guo, X., He, W., Ding, Y., Hong, B., Zhang, M., Wang, J., Jin, S., Zhou, E., et al. (2023) · arXiv · doi:10.48550/arxiv.2309.07864 · Cited by 972 · cited 6×
“According to [2], [7], there are generally five constructing modules of LM agents: planning, action, memory, interaction, and security.”
Influential
Research works printed but not linked: 164
- Shavit, 2023, Practices for governing agentic AI systems
- 10.1023/A:1015008417172 · DOI
- Paglieri, BALROG: Benchmarking agentic LLM and VLM reasoning on games, 1
- Cheng, 2024, Exploring large language model based intelligent agents: Definitions, methods, and prospects, arXiv:2401.03428 · arXiv
- 10.57020/ject.1297961 · DOI
- Wu, AutoGen: Enabling next-gen LLM applications via multi-agent conversation, 1
- 10.18653/v1/2024.findings-emnlp.270 · DOI
- Ahn, Do as I can, not as I say: Grounding language in robotic affordances, 1
- Nakano, 2022, WebGPT: Browser-assisted question-answering with human feedback, arXiv:2112.09332 · arXiv
- 10.52202/079017-0088 · DOI
- 10.1109/IROS58592.2024.10802229 · DOI
- Deng, PentestGPT: Evaluating and harnessing large language models for automated penetration testing, 847
- 10.1109/MWC.016.2300600 · DOI · Paper page
- Marro, 2025, Agora protocol: Scalable and reliable communication between your agents
- Zhang, VPGTrans: Transfer visual prompt generator across LLMs, 36, 20299
- Ma, LLM-pruner: On the structural pruning of large language models, 36, 21702
- 10.1609/aaai.v38i17.29860 · DOI
- 10.1109/MWC.005.2400019 · DOI
- Fedus, 2022, Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity, J. Mach. Learn. Res., 23, 1
- Zou, PoisonedRAG: Knowledge poisoning attacks to retrieval-augmented generation of large language models, 1
- Carlini, Quantifying Memorization across neural language models, 1
- Liu, 2023, Prompt injection attack against LLM-integrated applications, arXiv:2306.05499 · arXiv
- Fang, Local model poisoning attacks to Byzantine-robust federated learning, 1605
- Rando, Universal jailbreak backdoors from poisoned human feedback, 1
- Carlini, Extracting training data from large language models, 2633
- 10.1109/COMST.2024.3353265 · DOI · Paper page
- 10.24963/ijcai.2024/890 · doi:10.24963/ijcai.2024/890 · Paper page
- Durante, 2024, Agent AI: Surveying the horizons of multimodal interaction, arXiv:2401.03568 · arXiv
- 10.1109/comst.2025.3527641 · DOI · Paper page
- 10.1109/OJCOMS.2024.3456549 · DOI · Paper page
- Mei, 2024, AIOS: LLM agent operating system, arXiv:2403.16971 · arXiv
- Wei, Chain-of-thought prompting elicits reasoning in large language models, 35, 24824
- Yao, Tree of thoughts: Deliberate problem solving with large language models, 1
- Yao, ReAct: Synergizing reasoning and acting in language models, 1
- Shinn, Reflexion: Language agents with verbal reinforcement learning, 8634
- Wang, Voyager: An open-ended embodied agent with large language models, 1
- Hu, AVIS: Autonomous visual information seeking with large language model agent, 867
- Lewis, 2020, Retrieval-augmented generation for knowledge-intensive NLP tasks, Proc. NeurIPS, 33, 9459
- Kang, Knowledge-augmented reasoning distillation for small language models in knowledge-intensive tasks, 36, 1
- Liu, 2024, From LLM to conversational agent: A memory enhanced architecture with fine-tuning of large language models, arXiv:2401.02777 · arXiv
- Huang, Language models as zero-shot planners: Extracting actionable knowledge for embodied agents, 9118
- Driess, PaLM-E: An embodied multimodal language model, 202, 8469
- Talebirad, 2023, Multi-agent collaboration: Harnessing the power of intelligent LLM agents, arXiv:2306.03314 · arXiv
- Hong, MetaGPT: Meta programming for a multi-agent collaborative framework, 1
- 2025, Unitree G1
- 2025, Figure 02 humanoid robot
- 2025, Baidu Apollo launches Apollo ADFM large model
- 10.18653/v1/2023.emnlp-main.13 · DOI · Paper page
- Jinxin, 2023, CGMI: Configurable general multi-agent interaction framework, arXiv:2308.12503 · arXiv
- 10.1145/3637528.3671620 · DOI
- 10.1109/MNET.2024.3376419 · DOI
- Xu, 2024, AutoAttacker: A large language model guided system to implement automatic cyber-attacks, arXiv:2403.01038 · arXiv
- Guo, 2025, DeepSeek-R1: Incentivizing reasoning capability in LLMs via reinforcement learning, arXiv:2501.12948 · arXiv
- 2025, AgenticFlow AI
- 2025, Dify.AI
- 2025, RAGFlow
- 10.1109/MNET.2024.3435752 · DOI · Paper page
- Hao, ToolkenGPT: Augmenting frozen language models with massive tools via tool embeddings, 1
- 10.18653/v1/2023.findings-emnlp.691 · DOI · Paper page
- 10.1109/IROS58592.2024.10801334 · DOI
- Wan, Knowledge fusion of large language models, 1
- 10.1109/TIV.2024.3396450 · DOI
- 10.1609/aaai.v37i13.26928 · doi:10.1609/aaai.v37i13.26928
- 10.1109/TMC.2011.64 · doi:10.1109/tmc.2011.64
- 10.1109/mnet.2024.3442880 · DOI
- 10.1109/ICCV51070.2023.02137 · DOI
- 10.1109/comst.2024.3516819 · DOI
- 10.1145/1293731.1293735 · DOI · Paper page
- 10.1145/191246.191322 · DOI · Paper page
- 2024, Model context protocol (MCP)
- 2024, Agent network protocol (ANP)
- 10.1186/s43058-022-00380-5 · DOI
- 10.1109/tmc.2025.3546466 · DOI
- 10.1109/ICDCS60910.2024.00013 · DOI
- 10.18653/v1/2024.acl-long.363 · DOI
- Chan, ChatEval: Towards better LLM-based evaluators through multi-agent debate, 1
- Xu, 2024, Cached model-as-a-resource: Provisioning large language model agents for edge intelligence in space-air-ground integrated networks, arXiv:2403.05826 · arXiv
- Liu, A dynamic LLM-powered agent network for task-oriented agent collaboration, 1
- 10.1109/OJCS.2023.3300321 · DOI · Paper page
- Brown, Language models are few-shot learners, 33, 1877
- 10.1007/978-3-030-70604-3_4 · DOI
- Shen, Mixture-of-experts meets instruction tuning: A winning combination for large language models, 1
- 10.1525/9780520940420-020 · DOI · Paper page
- 10.1109/TWC.2022.3221778 · DOI
- 10.24963/ijcai.2024/3 · DOI
- 10.1109/TCNS.2023.3242359 · doi:10.1109/tcns.2023.3242359
- 10.1002/9781119863663.ch8 · DOI
- 10.1109/ICCV48922.2021.00148 · DOI
- 10.1609/aaai.v37i4.25555 · DOI · Paper page
- Qian, Scaling large language model-based multi-agent collaboration, 1
- Zhong, Seeking neural nuggets: Knowledge transfer in large language models from a parametric perspective, 1
- 10.18653/v1/2022.findings-naacl.26 · DOI
- 10.18653/v1/2022.findings-acl.220 · DOI
- 10.18653/v1/2023.acl-long.792 · DOI · Paper page
- Shen, Joint language semantic and structure embedding for knowledge graph completion, 1965
- 10.18653/v1/2024.findings-acl.400 · DOI
- 10.18653/v1/2020.emnlp-main.437 · DOI · Paper page
- 10.1109/TKDE.2023.3325484 · DOI
- Xu, An LLM can fool itself: A prompt-based adversarial attack, 1
- Du, Stable diffusion is unstable, 1
- 10.1109/SP54263.2024.00102 · DOI
- Luo, An image is worth 1000 lies: Transferability of adversarial images across prompts on vision-language models, 1
- 10.1109/SP54263.2024.00123 · DOI
- Deng, Jailbreaker: Automated jailbreak across multiple large language model chatbots, 1
- Yu, Don’t listen to me: Understanding and exploring jailbreak prompts of large language models, 1
- Zhang, BadRobot: Jailbreaking embodied LLMs in the physical world, 1
- Toyer, Tensor trust: Interpretable prompt injection attacks from an online game, 1
- Schuster, You autocomplete me: Poisoning vulnerabilities in neural code completion, 1559
- Zhang, Agent security bench (ASB): Formalizing and benchmarking attacks and defenses in LLM-based agents, 1
- 10.18653/v1/2025.emnlp-main.1771 · DOI
- Chan, 2024, AgentMonitor: A plug-and-play framework for predictive and secure multi-agent systems, arXiv:2408.14972 · arXiv
- 10.1109/ICCV51070.2023.00423 · DOI
- 10.18653/v1/2024.naacl-long.171 · DOI
- Shi, Large language models can be easily distracted by irrelevant context, 202, 31210
- Liu, Training socially aligned language models on simulated social interactions, 1
- Liang, Adversarial example does good: Preventing painting imitation from diffusion models via adversarial examples, 202, 20763
- 10.18653/v1/2023.acl-industry.56 · DOI
- 10.18653/v1/2020.acl-main.529 · DOI
- Kumar, 2023, Certifying LLM safety against adversarial prompting, arXiv:2309.02705 · arXiv
- Phute, LLM self defense: By self examination, LLMs know they are being tricked, 1
- Zeng, AutoDefense: Multi-agent LLM defense against jailbreak attacks, 1
- 10.14722/ndss.2024.24115 · DOI
- Jones, Automatically auditing large language models via discrete optimization, 15307
- Xu, 2024, RedAgent: Red teaming large language models with context-aware autonomous language agent, arXiv:2407.16667 · arXiv
- Zhou, WebArena: A realistic Web environment for building autonomous agents, 1
- 10.18653/v1/2024.findings-naacl.217 · DOI
- 10.18653/v1/2021.findings-emnlp.369 · DOI
- 10.14722/ndss.2024.23238 · DOI
- 10.18653/v1/2022.findings-emnlp.148 · DOI · Paper page
- 10.18653/v1/2023.acl-long.709 · DOI
- Carlini, Extracting training data from diffusion models, 5253
- 10.18653/v1/2022.emnlp-main.570 · DOI · Paper page
- Kong, An efficient membership inference attack for the diffusion model by proximal Initialization, 1
- 10.52202/079017-4290 · DOI
- Wang, Property existence inference against generative models, 1
- Krishna, Thieves on sesame street! model extraction of BERT-based APIs, 1
- 10.1145/3576915.3616652 · DOI
- 10.1145/3597503.3639091 · DOI
- 10.18653/v1/2023.emnlp-main.189 · DOI
- Shen, Prompt stealing attacks against text-to-image generation models, 1
- Sha, 2024, Prompt stealing attacks against large language models, arXiv:2402.12959 · arXiv
- Mauran, 2024, Samsung bans ChatGPT, AI Chatbots after data leak blunder
- 10.1109/JPROC.2023.3306773 · DOI · Paper page
- 10.1016/j.neucom.2021.04.105 · DOI
- 10.1109/SP.2019.00031 · DOI · Paper page
- Kandpal, User inference attacks on large language models, 1
- 10.18653/v1/2022.emnlp-main.119 · DOI · Paper page
- Kandpal, Deduplicating training data mitigates privacy risks in language models, 10697
- 10.18653/v1/2021.findings-emnlp.102 · DOI
- Kim, ProPILE: Probing privacy leakage in large language models, 1
- 10.1109/SP.2018.00038 · DOI · Paper page
- 10.1145/2810103.2813677 · DOI · Paper page
- Kirchenbauer, A watermark for large language models, 17061
- 10.1109/TSC.2024.3382958 · DOI
- Liu, Mitigating hallucination in large multi-modal models via robust instruction tuning, 1
- Penedo, The RefinedWeb dataset for falcon LLM: Outperforming curated corpora with Web data only, 1
- Lee, Factuality enhanced language models for open-ended text generation, 1
- Zhang, How language model hallucinations can snowball, 1
- Jones, Teaching language models to hallucinate less with synthetic tasks, 1
- Mündler, Self-contradictory hallucinations of large language models: Evaluation, detection and mitigation, 1
- Chen, INSIDE: LLMs’ internal states retain the power of hallucination detection, 1
- Zhou, Analyzing and mitigating object hallucination in large vision-language models, 1
- 10.1109/CVPR52729.2023.00373 · DOI
- 10.1109/ICCV48922.2021.00466 · DOI
Cited by
A Survey on Trustworthy LLM Agents: Threats and Countermeasures
Yu, M., Meng, F., Zhou, X., Wang, S., Mao, J., Pang, L., Chen, T., Wang, K., Li, X., Zhang, Y., et al. (2025) · arXiv · doi:10.48550/arxiv.2503.09648 · Cited by 34
A Survey on Trustworthy LLM Agents: Threats and Countermeasures
Yu, M., Meng, F., Zhou, X., Wang, S., Mao, J., Pan, L., Chen, T., Wang, K., Li, X., Zhang, Y., et al. (2025) · Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2 · doi:10.1145/3711896.3736561 · Cited by 34
Yan, W., Liu, Y.-L., Mamaeva, V., Dong, F., Tao, G., Li, R., & Yang, H. (2026) · Telecommunications Policy · doi:10.1016/j.telpol.2025.103117 · Cited by 28
Multi-Agent Debate Strategies: Survey, Taxonomy, and Challenges
Motger, Q., Oriol, M., Marco, J., & Franch, X. (2026) · arXiv · doi:10.48550/arxiv.2607.26212 · Cited by 5
Metadata
- Authors
- Wang, Y., Pan, Y., Su, Z., Deng, Y., Zhao, Q., Du, L., Luan, T. H., Kang, J., & Niyato, D.
- Venue
- IEEE Communications Surveys & Tutorials
- Published
- 2026
- Volume
- 28
- Pages
- 1906-1949
- ISSN
- 1553-877X (ISSN-L)
- 2373-745X
- OpenAlex
- W4410985905
- Cited by
- 68
- References
- 66
- FWCI
- 27.88
- Type
- journal article
- Language
- en
- Open access
- green
- Licence
- Funders
- NSFC (62302387, U22A2029, U24A20237, U23A20276, 62402379)
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.
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Large Model-Based Agents: State-of-the-Art, Cooperation Paradigms, Security and Privacy, and Future Trends. https://app.cochat.ai/papers/10.1109/comst.2025.3576176
- The paper
- Wang, Y., Pan, Y., Su, Z., Deng, Y., Zhao, Q., Du, L., Luan, T. H., Kang, J., & Niyato, D. (2026). Large Model-Based Agents: State-of-the-Art, Cooperation Paradigms, Security and Privacy, and Future Trends. IEEE Communications Surveys & Tutorials. https://doi.org/10.1109/comst.2025.3576176
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.