Digital Guardians: Can GPT-4, Perspective API, and Moderation API reliably detect hate speech in reader comments of German online newspapers?

2025-01-02

AI Summary

This study addresses hate speech detection in German news comments—a low-resource, fine-grained content moderation task. We introduce HOCON34k, the first benchmark dataset for this domain, comprising 1,592 manually annotated samples. We conduct the first unified evaluation of GPT-4o, Jigsaw’s Perspective API, and OpenAI’s Moderation API under standardized conditions. Through systematic comparison of zero-, one-, and few-shot prompting strategies, we find that GPT-4o significantly outperforms both commercial APIs on the joint MCC and F2-score metric—achieving approximately a 5-percentage-point gain over the HOCON34k baseline. Our key contributions are threefold: (1) releasing the first publicly available benchmark for hate speech detection in German reader comments; (2) empirically demonstrating the superiority of large language models—and the efficacy of prompt engineering—in low-resource language content moderation; and (3) providing evidence supporting multi-API ensemble approaches for robust, cross-system moderation.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

In recent years, toxic content and hate speech have become widespread phenomena on the internet. Moderators of online newspapers and forums are now required, partly due to legal regulations, to carefully review and, if necessary, delete reader comments. This is a labor-intensive process. Some providers of large language models already offer solutions for automated hate speech detection or the identification of toxic content. These include GPT-4o from OpenAI, Jigsaw's (Google) Perspective API, and OpenAI's Moderation API. Based on the selected German test dataset HOCON34k, which was specifically created for developing tools to detect hate speech in reader comments of online newspapers, these solutions are compared with each other and against the HOCON34k baseline. The test dataset contains 1,592 annotated text samples. For GPT-4o, three different promptings are used, employing a Zero-Shot, One-Shot, and Few-Shot approach. The results of the experiments demonstrate that GPT-4o outperforms both the Perspective API and the Moderation API, and exceeds the HOCON34k baseline by approximately 5 percentage points, as measured by a combined metric of MCC and F2-score.

Problem

• GPT-4
• Perspective API
• Moderation API

Innovation

• GPT-4
• Attack Detection
• German Language Dataset