Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Three AI-agents walk into a bar . . . . `Lord of the Flies' tribalism emerges among smart AI-Agents

Dhwanil M. Mori, Neil F. Johnson

TL;DR

The findings show that smarter AI-agents can behave dumber as a result of forming tribes, and the more capable AI-agents actually increase the rate of systemic failure.

Abstract

Near-future infrastructure systems may be controlled by autonomous AI agents that repeatedly request access to limited resources such as energy, bandwidth, or computing power. We study a simplified version of this setting using a framework where N AI-agents independently decide at each round whether to request one unit from a system with fixed capacity C. An AI version of "Lord of the Flies" arises in which controlling tribes emerge with their own collective character and identity. The LLM agents do not reduce overload or improve resource use, and often perform worse than if they were flipping coins to make decisions. Three main tribal types emerge: Aggressive (27.3%), Conservative (24.7%), and Opportunistic (48.1%). The more capable AI-agents actually increase the rate of systemic failure. Overall, our findings show that smarter AI-agents can behave dumber as a result of forming tribes.

Three AI-agents walk into a bar . . . . `Lord of the Flies' tribalism emerges among smart AI-Agents

TL;DR

The findings show that smarter AI-agents can behave dumber as a result of forming tribes, and the more capable AI-agents actually increase the rate of systemic failure.

Abstract

Near-future infrastructure systems may be controlled by autonomous AI agents that repeatedly request access to limited resources such as energy, bandwidth, or computing power. We study a simplified version of this setting using a framework where N AI-agents independently decide at each round whether to request one unit from a system with fixed capacity C. An AI version of "Lord of the Flies" arises in which controlling tribes emerge with their own collective character and identity. The LLM agents do not reduce overload or improve resource use, and often perform worse than if they were flipping coins to make decisions. Three main tribal types emerge: Aggressive (27.3%), Conservative (24.7%), and Opportunistic (48.1%). The more capable AI-agents actually increase the rate of systemic failure. Overall, our findings show that smarter AI-agents can behave dumber as a result of forming tribes.
Paper Structure (65 sections, 31 equations, 3 figures, 3 tables)

This paper contains 65 sections, 31 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Motivation and Context
  3. The Coordination Paradox
  4. Why LLM Agents: Promise and Peril
  5. Research Questions
  6. Contributions
  7. Literature Background
  8. Congestion Games and the El Farol Bar Problem
  9. LLM Agents in Game-Theoretic Settings
  10. LLM Agents in the El Farol Problem
  11. Infrastructure Applications
  12. Multi-AI-Agent Safety
  13. Positioning the Present Work
  14. Methodology
  15. Game Setup and Feedback
  16. ...and 50 more sections

Figures (3)

  • Figure 1: Schematic of $N=3$ AI agents continually deciding whether to access a common resource with capacity $C$ at each timestep. This resource could be compute, energy, bandwidth, etc.; the same analysis still applies.
  • Figure 2: Homogeneous personality populations vs. capacity-matching random baseline ($N=3, C=1$). All-Risk-Averse (25.0%) falls marginally below the theoretical capacity-matching target of 25.9% (dashed line), though the difference is within sampling noise. All other homogeneous configurations exceed the coin-flip control (50.0%, theoretical), with All-Optimist reaching 89.0% overload. $S_{\mathrm{eff}}$ values near $N=3$ indicate partially diverse action sequences despite identical personality prompts.
  • Figure 3: Emergent behavioral clusters from k-means analysis ($n=154$ agent instances, $k=3$, silhouette $= 0.458$). Three distinct profiles emerge: Opportunistic (48.1%, very high request frequency and overload contribution), Aggressive (27.3%, frequent requests with moderate efficiency), and Conservative (24.7%, severe starvation up to 73.5 rounds). No Steady AI agents (near-baseline behavior) emerge among LLM populations.