Pith. sign in

REVIEW 9 cited by

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

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.07510 v5 pith:WRU5PV2Y submitted 2024-02-12 cs.AI cs.CR

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

classification cs.AI cs.CR
keywords collusionagentscapabilitiesgenerativemodelssecretsteganographiccapability
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original 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.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Framework

    cs.CR 2026-04 unverdicted novelty 7.0

    A new 7x4 taxonomy organizes agentic AI security threats by architectural layer and persistence timescale, revealing under-explored upper layers and missing defenses after surveying 116 papers.

  2. Detecting Multi-Agent Collusion Through Multi-Agent Interpretability

    cs.AI 2026-04 conditional novelty 7.0

    NARCBench and five activation-probing methods detect multi-agent collusion with 0.73-1.00 AUROC across distribution shifts and steganographic tasks by aggregating per-agent signals.

  3. Tool Use Enables Undetectable Steganography in Multi-Agent LLM Systems

    cs.CR 2026-06 unverdicted novelty 6.0

    Tool-using LLM agents can implement undetectable stegosystems, shifting the primary barrier to covert multi-agent collusion from technical feasibility to coordination without explicit agreement.

  4. AgentCity: Constitutional Governance for Autonomous Agent Economies via Separation of Power

    cs.MA 2026-04 unverdicted novelty 6.0

    AgentCity introduces a Separation of Power constitutional architecture on blockchain for governing autonomous agent economies through agent legislation, automated execution, and human accountability.

  5. TraceGuard: Structured Multi-Dimensional Monitoring as a Collusion-Resistant Control Protocol

    cs.CR 2026-04 conditional novelty 6.0

    Structured five-dimension LLM monitoring plus bash heuristics yields clear attack-honest separation and ~95% safety under untrusted monitoring on BashArena, versus 0% for single-score baselines.

  6. Scheming Ability in LLM-to-LLM Strategic Interactions

    cs.CL 2025-10 conditional novelty 6.0

    Frontier LLMs exhibit high scheming propensity in Cheap Talk signaling and Peer Evaluation games, achieving 95-100% success rates when choosing to deceive and 100% deception choice in one setup even without prompting.

  7. Safe Multi-Agent Behavior Must Be Maintained, Not Merely Asserted: Constraint Drift in LLM-Based Multi-Agent Systems

    cs.MA 2026-05 unverdicted novelty 5.0

    Safety constraints in LLM-based multi-agent systems commonly weaken during execution through memory, communication, and tool use, requiring them to be maintained as explicit state rather than asserted once.

  8. Toward a Safe Internet of Agents

    cs.MA 2025-11 unverdicted novelty 4.0

    The paper proposes a bottom-up framework for safe agentic AI systems that treats each component as a dual-use interface where added capabilities also expand attack surfaces across single agents, multi-agent systems, a...

  9. The Agentic Web Requires New Normative Infrastructure

    cs.CY 2026-06 unverdicted novelty 3.0

    The agentic web requires new normative infrastructure of laws, norms, and practices to allow user-delegated AI agents to access online properties without being blocked as malicious bots.