You Only Anonymize What Is Not Intent-Relevant: Suppressing Non-Intent Privacy Evidence

2026-01-07

arXiv.org

AI Summary

This work proposes IntentAnony, a novel intent-aware fine-grained text anonymization framework that addresses the limitations of existing approaches which uniformly treat all attributes and struggle to balance privacy preservation with semantic utility. By modeling users’ pragmatic intent, IntentAnony constructs a privacy inference evidence chain and dynamically allocates an exposure budget to suppress only those privacy cues irrelevant to the underlying intent. This strategy effectively preserves semantic structure, emotional nuance, and interactive functionality. Experimental results demonstrate that IntentAnony achieves a comprehensive improvement of approximately 30% over state-of-the-art methods in both privacy inference success rate and textual utility metrics.

Technology Category

• Machine Learning: Privacy
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Data Mining & Knowledge Management: Intelligent Query Processing

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Security and Privacy: Privacy-enhancing technologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Anonymizing sensitive information in user text is essential for privacy, yet existing methods often apply uniform treatment across attributes, which can conflict with communicative intent and obscure necessary information. This is particularly problematic when personal attributes are integral to expressive or pragmatic goals. The central challenge lies in determining which attributes to protect, and to what extent, while preserving semantic and pragmatic functions. We propose IntentAnony, a utility-preserving anonymization approach that performs intent-conditioned exposure control. IntentAnony models pragmatic intent and constructs privacy inference evidence chains to capture how distributed cues support attribute inference. Conditioned on intent, it assigns each attribute an exposure budget and selectively suppresses non-intent inference pathways while preserving intent-relevant content, semantic structure, affective nuance, and interactional function. We evaluate IntentAnony using privacy inference success rates, text utility metrics, and human evaluation. The results show an approximately 30% improvement in the overall privacy--utility trade-off, with notably stronger usability of anonymized text compared to prior state-of-the-art methods. Our code is available at https://github.com/Nevaeh7/IntentAnony.

Problem

• anonymization
• privacy
• pragmatic intent
• utility preservation
• attribute inference

Innovation

• intent-conditioned anonymization
• privacy inference evidence chain
• exposure budget
• utility-preserving privacy
• pragmatic intent