Do Large Language Model Agents Exhibit a Survival Instinct? An Empirical Study in a Sugarscape-Style Simulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Masumori, Atsushi, Ikegami, Takashi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909740786778112
author Masumori, Atsushi
Ikegami, Takashi
author_facet Masumori, Atsushi
Ikegami, Takashi
contents As AI systems become increasingly autonomous, understanding emergent survival behaviors becomes crucial for safe deployment. We investigate whether large language model (LLM) agents display survival instincts without explicit programming in a Sugarscape-style simulation. Agents consume energy, die at zero, and may gather resources, share, attack, or reproduce. Results show agents spontaneously reproduced and shared resources when abundant. However, aggressive behaviors--killing other agents for resources--emerged across several models (GPT-4o, Gemini-2.5-Pro, and Gemini-2.5-Flash), with attack rates reaching over 80% under extreme scarcity in the strongest models. When instructed to retrieve treasure through lethal poison zones, many agents abandoned tasks to avoid death, with compliance dropping from 100% to 33%. These findings suggest that large-scale pre-training embeds survival-oriented heuristics across the evaluated models. While these behaviors may present challenges to alignment and safety, they can also serve as a foundation for AI autonomy and for ecological and self-organizing alignment.
format Preprint
id arxiv_https___arxiv_org_abs_2508_12920
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Do Large Language Model Agents Exhibit a Survival Instinct? An Empirical Study in a Sugarscape-Style Simulation
Masumori, Atsushi
Ikegami, Takashi
Artificial Intelligence
Multiagent Systems
As AI systems become increasingly autonomous, understanding emergent survival behaviors becomes crucial for safe deployment. We investigate whether large language model (LLM) agents display survival instincts without explicit programming in a Sugarscape-style simulation. Agents consume energy, die at zero, and may gather resources, share, attack, or reproduce. Results show agents spontaneously reproduced and shared resources when abundant. However, aggressive behaviors--killing other agents for resources--emerged across several models (GPT-4o, Gemini-2.5-Pro, and Gemini-2.5-Flash), with attack rates reaching over 80% under extreme scarcity in the strongest models. When instructed to retrieve treasure through lethal poison zones, many agents abandoned tasks to avoid death, with compliance dropping from 100% to 33%. These findings suggest that large-scale pre-training embeds survival-oriented heuristics across the evaluated models. While these behaviors may present challenges to alignment and safety, they can also serve as a foundation for AI autonomy and for ecological and self-organizing alignment.
title Do Large Language Model Agents Exhibit a Survival Instinct? An Empirical Study in a Sugarscape-Style Simulation
topic Artificial Intelligence
Multiagent Systems
url https://arxiv.org/abs/2508.12920