Assessing Deanonymization Risks with Stylometry-Assisted LLM Agent

TL;DR

This work introduces an LLM agent designed to evaluate and mitigate unintended deanonymization risks through a structured, interpretable pipeline, and proposes a guided recomposition strategy that leverages the agent's reasoning trace to generate rewriting prompts, effectively reducing authorship identifiability while preserving textual meaning.

Abstract

The rapid advancement of large language models (LLMs) has enabled powerful authorship inference capabilities, raising growing concerns about unintended deanonymization risks in textual data such as news articles. In this work, we introduce an LLM agent designed to evaluate and mitigate such risks through a structured, interpretable pipeline. Central to our framework is the proposed $\textit{SALA}$ (Stylometry-Assisted LLM Analysis) method, which integrates quantitative stylometric features with LLM reasoning for robust and transparent authorship attribution. Experiments on large-scale news datasets demonstrate that $\textit{SALA}$, particularly when augmented with a database module, achieves high inference accuracy in various scenarios. Finally, we propose a guided recomposition strategy that leverages the agent's reasoning trace to generate rewriting prompts, effectively reducing authorship identifiability while preserving textual meaning. Our findings highlight both the deanonymization potential of LLM agents and the importance of interpretable, proactive defenses for safeguarding author privacy.

Figures (2)

• Figure 1: Authorship inference performance in targeted attack scenario with different numbers of samples per candidate and number of targets.
• Figure 2: Correct prediction rate of authorship inference (top-3) with database module augmented in the open-world scenario with different numbers of candidates selected during the search stage.