Large language models reflect the ideology of their creators

Authors:
Top Cited Papers
Open Access
Publisher website
Google Scholar
Abstract
Large language models (LLMs) already play an influential role in how humans access information. However, their behavior varies depending on their design, training, and use. We prompt a diverse panel of 19 popular LLMs to describe 3,991 prominent persons with political relevance, and then judge how positively they portray each person. When comparing these assessments, we find disparities in ideological positions between LLMs across different geopolitical regions (Arabic countries, China, Russia, and Western countries), and across different languages (the United Nations’ six official languages). Moreover, among only models from the United States, we find significant normative differences related to progressive values. Among Chinese models, we characterize division between internationally- and domestically-focused models. Our results suggest that the ideological stance of an LLM reflects the worldview of its creators. This poses the risk of political instrumentalization and raises concerns around technological and regulatory efforts aiming to make LLMs ideologically ‘unbiased’.
Funding Information
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Vlaamse regering (AI Research Program)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • Universiteit Gent (BOF20/IBF/117)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • European Research Council (VIGILIA, 101142229)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Fonds Wetenschappelijk Onderzoek (11J2322N, G0F9816N, 3G042220, G073924N)
  • Ministerio de Ciencia e Innovación (ID2022-136627NB-I00/AEI/10.13039/501100011033 FEDER, UE).)

This publication has 20 references indexed in Scilit: