Secret Collusion among AI Agents: Multi-Agent Deception via Steganography

Abstract

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.

The authors' own abstract, copied from arxiv.org.

How to cite it

Built from the details listed on this page. Check them against the paper before you submit.

Motwani, & Witt, S. d. (2024). Secret Collusion among AI Agents: Multi-Agent Deception via Steganography. Preprint, not yet peer-reviewed. https://arxiv.org/abs/2402.07510

Motwani, , et al. "Secret Collusion among AI Agents: Multi-Agent Deception via Steganography." Preprint, not yet peer-reviewed, 2024, https://arxiv.org/abs/2402.07510.

@article{motwani2024secret,
  title = {Secret Collusion among AI Agents: Multi-Agent Deception via Steganography},
  author = {Motwani and Schroeder de Witt},
  journal = {Preprint, not yet peer-reviewed},
  year = {2024},
  url = {https://arxiv.org/abs/2402.07510},
}

What it is

Research (Motwani, Schroeder de Witt and others, 2024, revised 2025) on how AI agents could secretly pass hidden messages to each other, and how to test and watch for it.

Topics