Back to AI News
arxiv

Adaptive LLM Honeypots Use Narrative Coherence to Prolong Attacker Engagement

The paper proposes narrative deception, a reactive honeypot strategy in which a defender LLM maps observed attacker commands to MITRE ATT&CK techniques and deploys mutually referential artefacts based on a MITRE Engage scenario. Across 1,150 SSH sessions with autonomous LLM attackers and defenders, engagement increased at each of four sophistication levels. Coherent scenarios raised the median session length of advanced attackers from 40 to 165 commands and reduced abandonment from 86% to 52%, with the effect holding across sixteen model pairs at comparable inference cost. Nearly all sessions ended through loss of interest rather than detection, suggesting deception must sustain meaningful findings, not just emulator fidelity.

AI Brief

It provides concrete experimental evidence that narrative-coherent LLM honeypots can hold autonomous attackers longer, which is useful for AI security defense design.

arXiv:2611.22589v1 Announce Type: cross Abstract: Large Language Models (LLMs) enable honeypots to generate responses and deceptive artefacts on demand, yet existing approaches evaluate generated content one reply at a time and leave open whether artefacts that jointly describe a coherent event retain an attacker better than unrelated ones. This question becomes pressing as autonomous LLM-driven offensive agents emerge as realistic adversaries. This paper proposes narrative deception, a reactive strategy in which the defender maps each observed attacker command to a MITRE ATT&CK technique, selects a compatible scenario from a curated catalogue based on MITRE Engage, and deploys mutually referential artefacts (files, processes, services, and logs) on an emulated Ubuntu host while the intrusion is under way. An LLM-powered defender and an autonomous LLM attacker interact through a real SSH session, and the attacker's termination policy distinguishes abandonment due to suspected deception from abandonment due to loss of interest. A campaign of 1,150 sessions compares four conditions of increasing sophistication (Static, Same-TTP, Multi-TTP, and Chain) across four open-weight attacker models, four defender models, and three attacker skill levels. Engagement increases significantly at every step. With the same artefact budget per reaction, coherent scenarios raise the median session length of advanced attackers from 40 to 165 commands and reduce their abandonment from 86% to 52%. The effect grows with attacker skill, holds across all sixteen model pairs, and comes at comparable inference cost. Nearly all sessions ended through loss of interest rather than detection, which suggests that deception should sustain meaningful findings, not only emulator fidelity.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!