AI 智能体之间的秘密串通：通过隐写术实现的多智能体欺骗
Secret Collusion among AI Agents: Multi-Agent Deception via Steganography
AI总结：
本文形式化生成式 AI 智能体借助隐写术秘密串通的问题，提出缓解措施与评估框架，并通过多模型实验发现 GPT-4 隐写能力显著跃升。
摘要：
大型语言模型（LLM）近期的能力提升，开启了由相互通信的生成式 AI 智能体群体解决联合任务的应用场景。这给未经授权的信息共享或其他不受欢迎的智能体协调形式带来了隐私和安全挑战。现代隐写术技术可能使此类动态难以被检测。本文借鉴 AI 与安全文献中的相关概念，全面形式化了生成式 AI 智能体系统中的秘密串通问题。我们研究了使用隐写术的动机，并提出了多种缓解措施。我们的研究形成了一个模型评估框架，用于系统性测试各种秘密串通形式所需的能力。我们在一系列当代 LLM 上提供了大量实证结果。尽管当前模型的隐写能力仍然有限，但 GPT-4 表现出能力跃升，表明需要持续监测前沿模型的隐写能力。最后，我们通过制定一项全面的研究计划来缓解未来生成式 AI 模型之间串通的风险。
英文摘要：
Recent capability increases in large language models (LLMs) open up applications in which groups of communicating generative AI agents solve joint tasks. This poses privacy and security challenges concerning the unauthorised sharing of information, or other unwanted forms of agent coordination. Modern steganographic techniques could render such dynamics hard to detect. In this paper, we comprehensively formalise the problem of secret collusion in systems of generative AI agents by drawing on relevant concepts from both AI and security literature. We study incentives for the use of steganography, and propose a variety of mitigation measures. Our investigations result in a model evaluation framework that systematically tests capabilities required for various forms of secret collusion. We provide extensive empirical results across a range of contemporary LLMs. While the steganographic capabilities of current models remain limited, GPT-4 displays a capability jump suggesting the need for continuous monitoring of steganographic frontier model capabilities. We conclude by laying out a comprehensive research program to mitigate future risks of collusion between generative AI models.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
