Riccardo Vasellini | Searcharxiv

arXiv2026

Is Lying an Emergent Behaviour in LLMs? Evidence from Gaslighting AI agents in a Sustainability Game

LLMs agents are increasingly used in multi-agent settings, yet their behaviour in sustainability games remains largely unexplored. This work investigates whether lying can emerge among LLM agents in a competitive sustainability game in which agents are informed that common resources can regenerate, although regeneration does not actually occur. We develop an agent-based model of a sustainability game in which agents manage industrial, military, and ecological resources, and interact through a network. LLM agents can observe neighbours' status, declare future attacks, receive permission to lie, and access reputation information, while rule-based agents provide an interpretable behavioural baseline. The results show that neighbour information strongly changes system dynamics, increasing attacks while improving biosphere retention and coexistence. Also, the presence of future declarations reduce extinction risk without suppressing conflict. Behaviourally, deception emerges even when agents are not explicitly allowed to lie, and explicit permission mainly increases bluffing and diversion rather than direct backstabbing. Finally, the presence of reputation memory and information about the current biosphere level reduces system ecological depletion. These findings suggest that deception can arise as an emergent behaviour in LLM-agent systems and that communication between LLM-agents could support sustainability while dealing with risk.

cs.MA↗

arXiv2025

Cooperative Behavior in Pre-State Societies: An Agent-Based Approach of the Aksum Civilization

This study intends to test the hypothesis that, contrary to traditional interpretation, the social structure of the polity of Aksum - especially in its early stages - was not characterized by a vertical hierarchy with highly centralized administrative power, and that the leaders mentioned in the few available inscriptions were predominantly ritual leaders with religious rather than coercive political authority. This hypothesis, suggested by the available archaeological evidence, is grounded in Charles Stanish's model, which posits that pre-state societies could achieve cooperative behavior without the presence of coercive authority. Using agent-based modeling applied to data inspired by the Aksum civilization, we examine the dynamics of cooperation in the presence and absence of a Public Goods Game. Results show that while cooperative behavior can emerge in the short term without coercive power, it may not be sustainable over the long term, suggesting a need for centralized authority to foster stable, complex societies. These findings provide insights into the evolutionary pathways that lead to state formation and complex social structures.

physics.soc-ph↗