publications

2026

1. NDSS ’26

   There is No War in Ba Sing Se: A Global Analysis of Content Moderation in Large Language Models

   Friedemann Lipphardt, Moonis Ali, Anja Feldmann, and 1 more author

   In Network and Distributed System Security (NDSS) Symposium, Feb 2026

   Abs PDF

   First global examination of LLM content moderation across 15 models from 10 companies, 12 vantage points and 13 languages, evaluated over 700k queries. Soft-moderation rates differ by up to 60% relative across locations; political content shows the most geographic variability; and online and locally-deployed versions of the same model can diverge significantly.

2. FOCI ’26

   Dual Standards: Examining Content Moderation Disparities Between API and WebUI Interfaces in Large Language Models

   Friedemann Lipphardt, Moonis Ali, Anja Feldmann, and 1 more author

   In Free and Open Communications on the Internet (FOCI), Jan 2026

   Abs PDF

   First study evaluating differences in LLM content moderation when queried via API versus WebUI, providing evidence for interface-specific filtering pipelines.

3. Submission

   Rabid Rabbit-hole?: Safety Audit of TikTok’s Recommendation System

   Friedemann Lipphardt and others

   2026

   Under submission

   Abs

   First study of political and contextual recommendation bias in TikTok feeds targeting underage users in Brazil and the US.

2025

1. EuroS&P ’25

   Can You Hear Me? A First Study of VoIP Censorship Techniques in Saudi Arabia and the UAE

   Friedemann Lipphardt, Anja Feldmann, and Devashish Gosain

   In IEEE European Symposium on Security and Privacy (EuroS&P), Jun 2025

   Abs PDF

   First study of Voice-over-IP censorship in the Middle East, characterising the blocking mechanics against major calling services nationwide.

2. AINTEC ’25

   You wanna see some real speed? Comparative Analysis of global Measurement Lab and Cloudflare speed test results

   Friedemann Lipphardt, Malte Tashiro, and Romain Fontugne

   In Proceedings of the 20th Asian Internet Engineering Conference (AINTEC), Nov 2025

   Abs PDF

   Cross-service comparison of M-Lab and Cloudflare speedtest data at global scale, characterising measurement biases and outcome divergence.

3. FOCI ’25

   1-800-Censorship: Analyzing internet censorship data using the Internet Yellow Pages

   Friedemann Lipphardt, Malte Tashiro, and Romain Fontugne

   In Free and Open Communications on the Internet (FOCI), Jul 2025

   Abs PDF

   Integration of internet censorship measurement data into the Internet Yellow Pages graph and showcase of initial cross-dataset queries.