AgentStealth: Reinforcing Large Language Model for Anonymizing User-generated Text

TL;DR

AgentStealth introduces a self-reinforcing framework for anonymizing user-generated text using locally deployed language models, balancing privacy protection and utility without relying on cloud-based solutions.

Contribution

It proposes a novel self-reinforcing anonymization framework combining adversarial workflows, contrastive learning, and reinforcement learning for effective local text anonymization.

Findings

Outperforms baselines in anonymization effectiveness (+12.3%)

Enhances utility preservation (+6.8%)

Supports deployment on edge devices

Abstract

In today's digital world, casual user-generated content often contains subtle cues that may inadvertently expose sensitive personal attributes. Such risks underscore the growing importance of effective text anonymization to safeguard individual privacy. However, existing methods either rely on rigid replacements that damage utility or cloud-based LLMs that are costly and pose privacy risks. To address these issues, we explore the use of locally deployed smaller-scale language models (SLMs) for anonymization. Yet training effective SLMs remains challenging due to limited high-quality supervision. To address the challenge, we propose AgentStealth, a self-reinforcing LLM anonymization framework.First, we introduce an adversarial anonymization workflow enhanced by In-context Contrastive Learning and Adaptive Utility-Aware Control. Second, we perform supervised adaptation of SLMs using high-quality data collected from the workflow, which includes both anonymization and attack signals. Finally, we apply online reinforcement learning where the model leverages its internal adversarial feedback to iteratively improve anonymization performance. Experiments on two datasets show that our method outperforms baselines in both anonymization effectiveness (+12.3%) and utility (+6.8%). Our lightweight design supports direct deployment on edge devices, avoiding cloud reliance and communication-based privacy risks. Our code is open-source at https://github.com/tsinghua-fib-lab/AgentStealth.