Extended Abstract: CensorAlert – Leveraging LLM Agents for Automated Censorship Report Aggregation and Analysis
Authors: Ali Zohaib (University of Massachusetts Amherst), Jade Sheffey (University of Massachusetts Amherst), Mingshi Wu (GFW Report), Amir Houmansadr (University of Massachusetts Amherst)
Year: 2026
Issue: 1
Pages: 11–13
Abstract: Internet censorship reporting is fragmented across many channels, from measurement platforms to multilingual, crowdsourced reports circulating in chat groups, forums, and on social platforms, making it hard for researchers and advocates to reliably track and act on emerging incidents. In this work, we introduce CensorAlert, a platform that addresses this challenge by aggregating reports from diverse sources and using LLM-based agents to continuously monitor, normalize, translate, and summarize them into a unified format. Each report is scored by an LLM agent for significance and surfaced as part of a ranked feed published at https://censoralert.org. Users can also subscribe to receive timely alerts via email or a supported messaging platform.
Copyright in FOCI articles are held by their authors. This article is published under a Creative Commons Attribution 4.0 license.
