Cover Image

When artificial minds negotiate: dark personality and the ultimatum game in large language models

We investigate if Large Language Models (LLMs) exhibit personality-driven strategic behavior in the Ultimatum Game by manipulating Dark Factor of Personality (D-Factor) profiles via standardized prompts. Across 400k decisions from 17 open-source models and 4,166 human benchmarks, we test whether LLM...

Full description

Saved in:
Bibliographic Details
Main Authors: Ferraz, Vinícius (Author) , Oláh, Tamás (Author) , Sazedul, Ratin (Author) , Schmidt, Robert J. (Author) , Schwieren, Christiane (Author)
Format: Book/Monograph Working Paper
Language:English
Published: Heidelberg Heidelberg University, Department of Economics 16 Dez. 2025
Series:AWI discussion paper series no. 768 (November 2025)
In: AWI discussion paper series (no. 768 (November 2025))

DOI:10.11588/heidok.00037813
Subjects:
Graue Literatur
Online Access:Verlag, kostenfrei: https://archiv.ub.uni-heidelberg.de/volltextserver/37813/1/Ferraz_Olah_Sazedul_et._al._2025_dp768%20A.pdf
Resolving-System, kostenfrei: https://nbn-resolving.org/urn:nbn:de:bsz:16-heidok-378137
Resolving-System, kostenfrei: https://doi.org/10.11588/heidok.00037813
Get full text
Author Notes:Vinícius Ferraz, Tamas Olah, Ratin Sazedul, Robert Schmidt, Christiane Schwieren
Description
Summary:We investigate if Large Language Models (LLMs) exhibit personality-driven strategic behavior in the Ultimatum Game by manipulating Dark Factor of Personality (D-Factor) profiles via standardized prompts. Across 400k decisions from 17 open-source models and 4,166 human benchmarks, we test whether LLMs playing the proposer and responder roles exhibit systematic behavioral shifts across five D-Factor levels (from least to most selfish). The proposer role exhibited strong monotonic declines in fair offers from 91% (D1) to 17% (D5), mirroring human patterns but with 34% steeper gradients, indicating hypersensitivity to personality prompts. Responders diverged sharply: where humans became more punitive at higher D-levels, LLMs maintained high acceptance rates (75-92%) with weak or reversed D-Factor sensitivity, failing to reproduce reciprocity-punishment dynamics. These role-specific patterns align with strong-weak situation accounts-personality matters when incentives are ambiguous (proposers) but is muted when contingent (responders). Cross-model heterogeneity was substantial: Models exhibiting the closest alignment with human behavior, according to composite similarity scores (integrating prosocial rates, D-Factor correlations, and odds ratios), were dolphin3, deepseek_1.5b, and llama3.2 (0.74-0.85), while others exhibited extreme or non-variable behavior. Temperature settings (0.2 vs. 0.8) exerted minimal influence. We interpret these patterns as prompt-driven regularities rather than genuine motivational processes, suggesting LLMs can approximate but not fully replicate human strategic behavior in social dilemmas.
Physical Description:Online Resource
DOI:10.11588/heidok.00037813

Similar Items