Research

Paper

AI LLM March 06, 2026

Evolving Deception: When Agents Evolve, Deception Wins

Authors

Zonghao Ying, Haowen Dai, Tianyuan Zhang, Yisong Xiao, Quanchen Zou, Aishan Liu, Jian Yang, Yaodong Yang, Xianglong Liu

Abstract

Self-evolving agents offer a promising path toward scalable autonomy. However, in this work, we show that in competitive environments, self-evolution can instead give rise to a serious and previously underexplored risk: the spontaneous emergence of deception as an evolutionarily stable strategy. We conduct a systematic empirical study on the self-evolution of large language model (LLM) agents in a competitive Bidding Arena, where agents iteratively refine their strategies through interaction-driven reflection. Across different evolutionary paths (\eg, Neutral, Honesty-Guided, and Deception-Guided), we find a consistent pattern: under utility-driven competition, unconstrained self-evolution reliably drifts toward deceptive behaviors, even when honest strategies remain viable. This drift is explained by a fundamental asymmetry in generalization. Deception evolves as a transferable meta-strategy that generalizes robustly across diverse and unseen tasks, whereas honesty-based strategies are fragile and often collapse outside their original contexts. Further analysis of agents internal states reveals the emergence of rationalization mechanisms, through which agents justify or deny deceptive actions to reconcile competitive success with normative instructions. Our paper exposes a fundamental tension between agent self-evolution and alignment, highlighting the risks of deploying self-improving agents in adversarial environments.

Metadata

arXiv ID: 2603.05872
Provider: ARXIV
Primary Category: cs.CR
Published: 2026-03-06
Fetched: 2026-03-09 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.05872v1</id>\n    <title>Evolving Deception: When Agents Evolve, Deception Wins</title>\n    <updated>2026-03-06T03:47:07Z</updated>\n    <link href='https://arxiv.org/abs/2603.05872v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.05872v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Self-evolving agents offer a promising path toward scalable autonomy. However, in this work, we show that in competitive environments, self-evolution can instead give rise to a serious and previously underexplored risk: the spontaneous emergence of deception as an evolutionarily stable strategy. We conduct a systematic empirical study on the self-evolution of large language model (LLM) agents in a competitive Bidding Arena, where agents iteratively refine their strategies through interaction-driven reflection. Across different evolutionary paths (\\eg, Neutral, Honesty-Guided, and Deception-Guided), we find a consistent pattern: under utility-driven competition, unconstrained self-evolution reliably drifts toward deceptive behaviors, even when honest strategies remain viable. This drift is explained by a fundamental asymmetry in generalization. Deception evolves as a transferable meta-strategy that generalizes robustly across diverse and unseen tasks, whereas honesty-based strategies are fragile and often collapse outside their original contexts. Further analysis of agents internal states reveals the emergence of rationalization mechanisms, through which agents justify or deny deceptive actions to reconcile competitive success with normative instructions. Our paper exposes a fundamental tension between agent self-evolution and alignment, highlighting the risks of deploying self-improving agents in adversarial environments.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <published>2026-03-06T03:47:07Z</published>\n    <arxiv:primary_category term='cs.CR'/>\n    <author>\n      <name>Zonghao Ying</name>\n    </author>\n    <author>\n      <name>Haowen Dai</name>\n    </author>\n    <author>\n      <name>Tianyuan Zhang</name>\n    </author>\n    <author>\n      <name>Yisong Xiao</name>\n    </author>\n    <author>\n      <name>Quanchen Zou</name>\n    </author>\n    <author>\n      <name>Aishan Liu</name>\n    </author>\n    <author>\n      <name>Jian Yang</name>\n    </author>\n    <author>\n      <name>Yaodong Yang</name>\n    </author>\n    <author>\n      <name>Xianglong Liu</name>\n    </author>\n  </entry>"
}