GEO-Flag: Detecting and Measuring GEO-Optimized Web Content

2026-08-17

AI Summary

This study addresses the proliferation of Generative Engine Optimization (GEO) content and the lack of auditing mechanisms in generative search. We construct a benchmark dataset and propose an Interventional Paired Training (IPT) strategy alongside a GEO-gated Agent auditing system. By leveraging ModernBERT fine-tuning and agent-based provenance tracing, we achieve precise GEO identification and citation verification. Experiments demonstrate a detection F1-score of 0.944 and improved long-tail accuracy of 0.883, while quantifying real-world GEO prevalence at 8.90% for the first time. This work bridges the gap in systematic detection, providing critical empirical evidence and technical support for governing the generative search ecosystem.

Abstract

Generative Engine Optimization (GEO) modifies web content to increase its likelihood of being selected and cited by generative search engines. This can give strategically optimized pages visibility disproportionate to their authority or relevance and even make weak or false information appear well supported. Unlike conventional search, generative search synthesizes information into direct answers rather than presenting competing sources, which can further amplify these risks, as assessing source provenance and authority requires additional user interaction. Despite these concerns, systematic methods for detecting GEO-optimized webpages remain underexplored. We introduce \texttt{GEOFlagBench}, a benchmark of 3,200 webpages spanning 400 queries, four domains, and eight GEO optimizer families, and use it to systematically evaluate existing GEO detection methods. Although the strongest baseline achieves an aggregate F1 of 0.880, method-level and authorship-conditioned evaluations reveal substantial weaknesses and potential reliance on authorship-related shortcuts. We therefore propose \emph{Intervention-Paired Training} (IPT), which supervises detector responses to GEO interventions and non-GEO AI polishing; on ModernBERT, IPT improves F1 from 0.862 to 0.944 and worst-group accuracy from 0.725 to 0.883. We develop a GEO-gated Agent system for auditing the Source Tier and verifiability of Citation URLs in detected GEO pages. Finally, we deploy the complete pipeline on released Google Search and Gemini-grounded retrieval results for 1,000 real-user queries. Across 10,095 available pages, we estimate an overall GEO prevalence of 8.90\%, reaching 16.36\% among pages modified in 2026. Our results establish a foundation for systematically detecting, auditing, and measuring GEO in real-world search ecosystems.

Problem

• Generative Engine Optimization
• GEO detection
• web content integrity
• generative search
• information provenance

Innovation

• Intervention-Paired Training
• GEOFlagBench
• GEO-gated Agent
• Generative Engine Optimization Detection