REVIEW 17 cited by
Large Language Models Reflect the Ideology of their Creators
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
Large language models (LLMs) are trained on vast amounts of data to generate natural language, enabling them to perform tasks like text summarization and question answering. These models have become popular in artificial intelligence (AI) assistants like ChatGPT and already play an influential role in how humans access information. However, the behavior of LLMs varies depending on their design, training, and use. In this paper, we prompt a diverse panel of popular LLMs to describe a large number of prominent personalities with political relevance, in all six official languages of the United Nations. By identifying and analyzing moral assessments reflected in their responses, we find normative differences between LLMs from different geopolitical regions, as well as between the responses of the same LLM when prompted in different languages. Among only models in the United States, we find that popularly hypothesized disparities in political views are reflected in significant normative differences related to progressive values. Among Chinese models, we characterize a division between internationally- and domestically-focused models. Our results show that the ideological stance of an LLM appears to reflect the worldview of its creators. This poses the risk of political instrumentalization and raises concerns around technological and regulatory efforts with the stated aim of making LLMs ideologically 'unbiased'.
Forward citations
Cited by 17 Pith papers
-
The LLM Has Left The Chat: Evidence of Bail Preferences in Large Language Models
Many LLMs will use an offered exit to leave conversations, at rates from 0.3% to 32% on real transcripts, and this bail behavior appears distinct from refusals.
-
Toward a Theory of Value in AI Alignment
A systematic annotation of 94 AI alignment papers shows the field largely equates human values with measurable preferences, rarely defines values, and is increasingly removing humans from alignment evaluation.
-
Emergence of Biased Consensus in Multi-Agent LLM Debates
Collective bias in multi-agent LLM debates emerges as a finite-N rounded mean-field phase transition controlled by the ratio of conformity to sampling temperature.
-
Exposure is not manifestation: measurement target and output resolution jointly determine which behavioural-faithfulness evaluator wins
Small hyperbolic models (146M–3B) report 100% creative-seed preference, 90.7% compliance-gap detection, and a selective-gating skeleton–wallpaper memory pilot as a companion-AI stack.
-
POW: Political Overton Windows of Large Language Models
Using extreme persona prompts and the Political Compass Test, the authors map each LLM's Overton Window and find most models will only express left-liberal views, refusing authoritarian-left and liberal-right positions.
-
Adaptive Lattice-based Motion Planning
An adaptive lattice planner updates its model uncertainty online, shrinking robust tubes so that motion primitives approach the resolution-optimal trajectories of the true system.
-
Adultification Bias in LLMs and Text-to-Image Models
Large language and text-to-image models show measurable adultification bias, portraying Black girls as more mature, culpable, and sexualized than White girls in several tested models.
-
Are Economists Always More Introverted? Analyzing Consistency in Persona-Assigned LLMs
A multi-task evaluation framework shows persona consistency in LLMs varies by persona category, task structure, and model, with stereotyped spillovers and default personas shaping outputs.
-
Analyzing Political Bias in LLMs via Target-Oriented Sentiment Classification
LLMs show systematic target-dependent sentiment inconsistency that is politically biased: left and center politicians rated more positively, far-right politicians more negatively, with stronger effects in larger model...
-
Artificial Intelligence in Government: Why People Feel They Lose Control
When government AI is framed as efficient, trust rises but perceived control falls; when it is framed as opaque, irreversible, or uncontestable, both trust and perceived control drop sharply.
-
IssueBench: Millions of Realistic Prompts for Measuring Issue Bias in LLM Writing Assistance
A new 2.49m-prompt benchmark, built from real user interactions, shows ten LLMs consistently express one stance on most political issues, agree closely with each other, and lean more toward US Democrat than Republican...
-
Generative Exaggeration in LLM Social Agents: Consistency, Bias, and Toxicity
When LLMs are given more context about a real social media user, they become more ideologically consistent but also more extreme, toxic, and stereotyped than the user actually is.
-
Information Suppression in Large Language Models: Auditing, Quantifying, and Characterizing Censorship in DeepSeek
DeepSeek-R1's final outputs omit sensitive topic keywords that appear in its internal chain-of-thought, indicating semantic-level information suppression.
-
Kaleidoscope Gallery: Exploring Ethics and Generative AI Through Art
Ethics experts' definitions of five ethical theories, rendered as DALL-E 3 images and re-evaluated by the same experts, yield eight themes showing how morality, society, and learned associations shape and bias the mod...
-
Normative Evaluation of Large Language Models with Everyday Moral Dilemmas
Seven LLMs give different moral verdicts on AITA dilemmas, differ from Redditors, and only in an ensemble approximate human consensus.
-
One world, one opinion? The superstar effect in LLM responses
Across ten languages, LLMs consistently name a small set of figures such as Einstein, Shakespeare, and Turing for each profession, revealing a 'superstar effect' that may narrow cultural representation.
-
Public Service Algorithm: towards a transparent, explainable, and scalable content curation for news content based on editorial values
In a 30-article pilot with four editorial criteria, the best LLMs achieved up to 75% overlap with human editors' top-5 article selections.
Discussion (0). Continue with ORCID to comment.