Search papers, labs, and topics across Lattice.
This paper introduces a multi-agent simulation in a simplified NYC model to study strategic behavior in LLM agents with opposing incentives: Blue agents navigate efficiently, while Red agents divert them to billboard routes using persuasive language. They iteratively update agent policies using Kahneman-Tversky Optimization (KTO), optimizing Blue agents for reduced billboard exposure and Red agents to exploit weaknesses. Results show that while Blue agents improve task success from 46.0% to 57.3%, they remain highly susceptible to adversarial persuasion (70.7%), revealing a persistent safety-helpfulness trade-off.
LLM agents in a simulated NYC learn to selectively trust and deceive, but remain surprisingly vulnerable to adversarial steering, highlighting a fundamental safety-helpfulness trade-off.
As large language models (LLMs) are increasingly deployed as autonomous agents, understanding how strategic behavior emerges in multi-agent environments has become an important alignment challenge. We take a neutral empirical stance and construct a controlled environment in which strategic behavior can be directly observed and measured. We introduce a large-scale multi-agent simulation in a simplified model of New York City, where LLM-driven agents interact under opposing incentives. Blue agents aim to reach their destinations efficiently, while Red agents attempt to divert them toward billboard-heavy routes using persuasive language to maximize advertising revenue. Hidden identities make navigation socially mediated, forcing agents to decide when to trust or deceive. We study policy learning through an iterative simulation pipeline that updates agent policies across repeated interaction rounds using Kahneman-Tversky Optimization (KTO). Blue agents are optimized to reduce billboard exposure while preserving navigation efficiency, whereas Red agents adapt to exploit remaining weaknesses. Across iterations, the best Blue policy improves task success from 46.0% to 57.3%, although susceptibility remains high at 70.7%. Later policies exhibit stronger selective cooperation while preserving trajectory efficiency. However, a persistent safety-helpfulness trade-off remains: policies that better resist adversarial steering do not simultaneously maximize task completion. Overall, our results show that LLM agents can exhibit limited strategic behavior, including selective trust and deception, while remaining highly vulnerable to adversarial persuasion.
