Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes

The paper aims to address large language models’ (LLMs) cultural bias using the World Value Survey Wave 7 (WVS) questionnaire on social values, attitudes, and stereotypes. Comparative analysis and LLMs interview methods measure the Euclidean distance of response vectors of four culturally diverse LL...

Full description

Saved in:
Bibliographic Details
Main Authors: Kristian Dokic, Barbara Pisker, Bojan Radisic
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Societies
Subjects:
Online Access:https://www.mdpi.com/2075-4698/15/5/142
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849718758815301632
author Kristian Dokic
Barbara Pisker
Bojan Radisic
author_facet Kristian Dokic
Barbara Pisker
Bojan Radisic
author_sort Kristian Dokic
collection DOAJ
description The paper aims to address large language models’ (LLMs) cultural bias using the World Value Survey Wave 7 (WVS) questionnaire on social values, attitudes, and stereotypes. Comparative analysis and LLMs interview methods measure the Euclidean distance of response vectors of four culturally diverse LLMs (USA, China, Russia, UAE) in a multidimensional vector space to contrast originated WVS research countries and population positions. The results confirmed the initial hypotheses reflecting culturally and linguistically biased LLM answers, considering specific socio-cultural contexts and English language and Latin script digital dominance in available training materials. USA-constructed LLMs showed the most liberal attitudes, followed by China, Russia, and the UAE. LLM interview results also show WVS results closest to the United States population, positioning the similarity of the responses in first place for China and Russia followed by the USA and the UAE. Mitigating initiatives in LLMs’ cultural and linguistic debiasing is required to preserve cultural and linguistic diversity in the digital space.
format Article
id doaj-art-e96fcd50e7914fdca30252ce039eb8cc
institution DOAJ
issn 2075-4698
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Societies
spelling doaj-art-e96fcd50e7914fdca30252ce039eb8cc2025-08-20T03:12:18ZengMDPI AGSocieties2075-46982025-05-0115514210.3390/soc15050142Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and StereotypesKristian Dokic0Barbara Pisker1Bojan Radisic2Faculty of Tourism and Rural Development, University of Osijek, 34000 Pozega, CroatiaFaculty of Tourism and Rural Development, University of Osijek, 34000 Pozega, CroatiaFaculty of Tourism and Rural Development, University of Osijek, 34000 Pozega, CroatiaThe paper aims to address large language models’ (LLMs) cultural bias using the World Value Survey Wave 7 (WVS) questionnaire on social values, attitudes, and stereotypes. Comparative analysis and LLMs interview methods measure the Euclidean distance of response vectors of four culturally diverse LLMs (USA, China, Russia, UAE) in a multidimensional vector space to contrast originated WVS research countries and population positions. The results confirmed the initial hypotheses reflecting culturally and linguistically biased LLM answers, considering specific socio-cultural contexts and English language and Latin script digital dominance in available training materials. USA-constructed LLMs showed the most liberal attitudes, followed by China, Russia, and the UAE. LLM interview results also show WVS results closest to the United States population, positioning the similarity of the responses in first place for China and Russia followed by the USA and the UAE. Mitigating initiatives in LLMs’ cultural and linguistic debiasing is required to preserve cultural and linguistic diversity in the digital space.https://www.mdpi.com/2075-4698/15/5/142large language modelculturebiaslanguagesociety
spellingShingle Kristian Dokic
Barbara Pisker
Bojan Radisic
Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
Societies
large language model
culture
bias
language
society
title Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
title_full Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
title_fullStr Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
title_full_unstemmed Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
title_short Mirroring Cultural Dominance: Disclosing Large Language Models Social Values, Attitudes and Stereotypes
title_sort mirroring cultural dominance disclosing large language models social values attitudes and stereotypes
topic large language model
culture
bias
language
society
url https://www.mdpi.com/2075-4698/15/5/142
work_keys_str_mv AT kristiandokic mirroringculturaldominancedisclosinglargelanguagemodelssocialvaluesattitudesandstereotypes
AT barbarapisker mirroringculturaldominancedisclosinglargelanguagemodelssocialvaluesattitudesandstereotypes
AT bojanradisic mirroringculturaldominancedisclosinglargelanguagemodelssocialvaluesattitudesandstereotypes