This study addresses security and ethical challenges in LLM-based Multi-Agent Systems, as exemplified in a blockchain fraud detection case study. Leveraging blockchain’s secure architecture, the framework involves specialized LLM Agents—ContractMining, Investigative, Ethics, and PerformanceMonitor, coordinated by a ManagerAgent. Baseline LLM models achieved 30% accuracy with a threshold method and 94% accuracy with a random-forest method. The Claude 3.5-powered LLM system reached an accuracy of 92%. Ethical evaluations revealed biases, highlighting the need for fairness-focused refinements. Our approach aims to develop trustworthy and reliable networks of agents capable of functioning even in adversarial environments. To our knowledge, no existing systems employ ethical LLM agents specifically designed to detect fraud, making this a novel contribution. Future work will focus on refining ethical frameworks, scaling the system, and benchmarking it against traditional methods to establish a robust, adaptable, and ethically grounded solution for blockchain fraud detection.
References
[1]
Deng, Z., Guo, Y., Han, C., Ma, W., Xiong, J., Wen, S. and Xiang, Y. (2024) AI Agents under Threat: A Survey of Key Security Challenges and Future Pathways. arXiv: 2406.02630.
[2]
Humphreys, D., Koay, A., Desmond, D. and Mealy, E. (2024) AI Hype as a Cyber Security Risk: The Moral Responsibility of Implementing Generative AI in Business. AIandEthics, 4, 791-804. https://doi.org/10.1007/s43681-024-00443-4
[3]
Wang, J., Hong, Y., Wang, J., Xu, J., Tang, Y., Han, Q., et al. (2022) Cooperative and Competitive Multi-Agent Systems: From Optimization to Games. IEEE/CAAJournalofAutomaticaSinica, 9, 763-783. https://doi.org/10.1109/jas.2022.105506
[4]
Motwani, S.R., Baranchuk, M., Hammond, L. and de Witt, C.S. (2023) A Perfect Collusion Benchmark: How Can AI Agents Be Prevented from Colluding with Information-Theoretic Undetectability? Multi-Agent Security Workshop Neu-rIPS’23.
[5]
Zeng, Y., Wu, Y., Zhang, X., Wang, H. and Wu, Q. (2024) Autodefense: Multi-Agent LLM Defense against Jailbreak Attacks. arXiv: 2403.04783.
[6]
Calvaresi, D., Calbimonte, J., Dubovitskaya, A., Mattioli, V., Piguet, J. and Schumacher, M. (2019) The Good, the Bad, and the Ethical Implications of Bridging Blockchain and Multi-Agent Systems. Information, 10, Article 363. https://doi.org/10.3390/info10120363
[7]
Woodward, C.R. (2022) Analysis of Integrating Blockchain Technologies into Multi-Agent Systems. arXiv: 2212.12313.
[8]
Zhuang, Y., Liu, Z., Qian, P., Liu, Q., Wang, X. and He, Q. (2020) Smart Contract Vulnerability Detection Using Graph Neural Network. Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, Yokohama, 11-17 July 2020, 3283-3290. https://doi.org/10.24963/ijcai.2020/454
[9]
Liu, L., Tsai, W., Bhuiyan, M.Z.A., Peng, H. and Liu, M. (2022) Blockchain-Enabled Fraud Discovery through Abnormal Smart Contract Detection on Ethereum. FutureGenerationComputerSystems, 128, 158-166. https://doi.org/10.1016/j.future.2021.08.023
[10]
Ravuri, A., Sendil, M.S., Rani, M., Srikanth, A., Sharath, M.N., Sudarsa, D., et al. (2024) Blockchain-Enabled Collaborative Anomaly Detection for IoT Security. MATECWebofConferences, 392, Article ID: 01141. https://doi.org/10.1051/matecconf/202439201141
[11]
Papi, F.G., Hübner, J.F. and de Brito, M. (2022) A Blockchain Integration to Support Transactions of Assets in Multi-Agent Systems. EngineeringApplicationsofArtificialIntelligence, 107, Article ID: 104534. https://doi.org/10.1016/j.engappai.2021.104534
[12]
Bajaj, S. (2023) Autonomous AI Agents in Fraud and Risk Management. https://oscilar.com/blog/autonomous-ai-agents
[13]
Park, T. (2024) Enhancing Anomaly Detection in Financial Markets with an LLM-based Multi-Agent Framework. arXiv: 2403.19735.
[14]
Gabriel, I., Manzini, A., Keeling, G., Hendricks, L.A., Rieser, V., Iqbal, H., Manyika, J., et al. (2024) The Ethics of Advanced AI Assistants. arXiv: 2404.16244.
[15]
Raji, I.D., Xu, P., Honigsberg, C. and Ho, D. (2022) Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance. Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society, Oxford, 19-21 May 2021, 557-571. https://doi.org/10.1145/3514094.3534181
[16]
Murikah, W., Nthenge, J.K. and Musyoka, F.M. (2024) Bias and Ethics of AI Systems Applied in Auditing—A Systematic Review. ScientificAfrican, 25, e02281. https://doi.org/10.1016/j.sciaf.2024.e02281
[17]
Jung, E., Le Tilly, M., Gehani, A. and Ge, Y. (2019) Data Mining-Based Ethereum Fraud Detection. 2019 IEEE International Conference on Blockchain (Blockchain), Atlanta, 14-17 July 2019, 266-273. https://doi.org/10.1109/blockchain.2019.00042
[18]
Hardt, M., Price, E. and Srebro, N. (2016) Equality of Opportunity in Supervised Learning. Proceedings of the 30th International Conference on Neural Information Processing Systems, 3323-3331.
[19]
Kleinberg, J., Mullainathan, S. and Raghavan, M. (2016) Inherent Trade-Offs in the Fair Determination of Risk Scores. arXiv: 1609.05807.
[20]
Barocas, S., Hardt, M. and Narayanan, A. (2023) Fairness and Machine Learning: Limitations and Opportunities. MIT Press.
