Longitudinal Monitoring of LLM Content Moderation of Social Issues

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Dai, Yunlang, Lurie, Emma, Metaxa, Danaé, Friedler, Sorelle A.
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Computers and Society Human-Computer Interaction
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866914071260954624
author	Dai, Yunlang Lurie, Emma Metaxa, Danaé Friedler, Sorelle A.
author_facet	Dai, Yunlang Lurie, Emma Metaxa, Danaé Friedler, Sorelle A.
contents	Large language models' (LLMs') outputs are shaped by opaque and frequently-changing company content moderation policies and practices. LLM moderation often takes the form of refusal; models' refusal to produce text about certain topics both reflects company policy and subtly shapes public discourse. We introduce AI Watchman, a longitudinal auditing system to publicly measure and track LLM refusals over time, to provide transparency into an important and black-box aspect of LLMs. Using a dataset of over 400 social issues, we audit Open AI's moderation endpoint, GPT-4.1, and GPT-5, and DeepSeek (both in English and Chinese). We find evidence that changes in company policies, even those not publicly announced, can be detected by AI Watchman, and identify company- and model-specific differences in content moderation. We also qualitatively analyze and categorize different forms of refusal. This work contributes evidence for the value of longitudinal auditing of LLMs, and AI Watchman, one system for doing so.
format	Preprint
id	arxiv_https___arxiv_org_abs_2510_01255
institution	arXiv
publishDate	2025
record_format	arxiv
spellingShingle	Longitudinal Monitoring of LLM Content Moderation of Social Issues Dai, Yunlang Lurie, Emma Metaxa, Danaé Friedler, Sorelle A. Computation and Language Computers and Society Human-Computer Interaction Large language models' (LLMs') outputs are shaped by opaque and frequently-changing company content moderation policies and practices. LLM moderation often takes the form of refusal; models' refusal to produce text about certain topics both reflects company policy and subtly shapes public discourse. We introduce AI Watchman, a longitudinal auditing system to publicly measure and track LLM refusals over time, to provide transparency into an important and black-box aspect of LLMs. Using a dataset of over 400 social issues, we audit Open AI's moderation endpoint, GPT-4.1, and GPT-5, and DeepSeek (both in English and Chinese). We find evidence that changes in company policies, even those not publicly announced, can be detected by AI Watchman, and identify company- and model-specific differences in content moderation. We also qualitatively analyze and categorize different forms of refusal. This work contributes evidence for the value of longitudinal auditing of LLMs, and AI Watchman, one system for doing so.
title	Longitudinal Monitoring of LLM Content Moderation of Social Issues
topic	Computation and Language Computers and Society Human-Computer Interaction
url	https://arxiv.org/abs/2510.01255