Longitudinal Monitoring of LLM Content Moderation of Social Issues
This work addresses the need for transparency in LLM moderation for researchers and the public, though it is incremental as it builds on existing auditing approaches.
The researchers tackled the problem of opaque and changing content moderation in large language models by developing AI Watchman, a longitudinal auditing system that tracks LLM refusals over time, finding evidence that it can detect unannounced policy changes and identify differences across models and companies.
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.