@inproceedings{huang-etal-2026-need,
title = "Need to Know: Contextual-Integrity-Grounded Query Rewriting for Privacy-Conscious {LLM} Delegation",
author = "Huang, Xinyue and
Yang, Wenyuan and
Cao, Xiaochun",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.300/",
pages = "6904--6919",
ISBN = "979-8-89176-459-0",
abstract = "As LLMs become increasingly woven into everyday workflows, user queries send to cloud-hosted LLMs routinely mix task-essential content with task-non-essential sensitive disclosures, yet type-based PII redaction is context-agnostic and may raise two issues: over disclosing untyped sensitive context and over-removing answer-bearing spans. We recast privacy-preserving query rewriting under Contextual Integrity: a span should be forwarded only if it is necessary for the task. We build the first task-based CI benchmark for privacy conscious delegation, comprising 3,167 samples that combine high-quality synthetic data spanning 11 tasks and 20 task types, WildChat based real-user queries, and a medical challenge set with dense sensitive information. Building on this benchmark, we propose a CI guided reinforcement learning framework that converts essential and non-essential sensitive spans into verifiable optimization signals, and train a query rewriter to preserve task-critical information while suppressing unnecessary sensitive disclosure. Experiments show that our learned rewriter achieves the best privacy{--}utility trade-off, achieving up to +10.1 average utility over on-device baselines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="huang-etal-2026-need">
<titleInfo>
<title>Need to Know: Contextual-Integrity-Grounded Query Rewriting for Privacy-Conscious LLM Delegation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xinyue</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenyuan</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaochun</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>As LLMs become increasingly woven into everyday workflows, user queries send to cloud-hosted LLMs routinely mix task-essential content with task-non-essential sensitive disclosures, yet type-based PII redaction is context-agnostic and may raise two issues: over disclosing untyped sensitive context and over-removing answer-bearing spans. We recast privacy-preserving query rewriting under Contextual Integrity: a span should be forwarded only if it is necessary for the task. We build the first task-based CI benchmark for privacy conscious delegation, comprising 3,167 samples that combine high-quality synthetic data spanning 11 tasks and 20 task types, WildChat based real-user queries, and a medical challenge set with dense sensitive information. Building on this benchmark, we propose a CI guided reinforcement learning framework that converts essential and non-essential sensitive spans into verifiable optimization signals, and train a query rewriter to preserve task-critical information while suppressing unnecessary sensitive disclosure. Experiments show that our learned rewriter achieves the best privacy–utility trade-off, achieving up to +10.1 average utility over on-device baselines.</abstract>
<identifier type="citekey">huang-etal-2026-need</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.300/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>6904</start>
<end>6919</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Need to Know: Contextual-Integrity-Grounded Query Rewriting for Privacy-Conscious LLM Delegation
%A Huang, Xinyue
%A Yang, Wenyuan
%A Cao, Xiaochun
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F huang-etal-2026-need
%X As LLMs become increasingly woven into everyday workflows, user queries send to cloud-hosted LLMs routinely mix task-essential content with task-non-essential sensitive disclosures, yet type-based PII redaction is context-agnostic and may raise two issues: over disclosing untyped sensitive context and over-removing answer-bearing spans. We recast privacy-preserving query rewriting under Contextual Integrity: a span should be forwarded only if it is necessary for the task. We build the first task-based CI benchmark for privacy conscious delegation, comprising 3,167 samples that combine high-quality synthetic data spanning 11 tasks and 20 task types, WildChat based real-user queries, and a medical challenge set with dense sensitive information. Building on this benchmark, we propose a CI guided reinforcement learning framework that converts essential and non-essential sensitive spans into verifiable optimization signals, and train a query rewriter to preserve task-critical information while suppressing unnecessary sensitive disclosure. Experiments show that our learned rewriter achieves the best privacy–utility trade-off, achieving up to +10.1 average utility over on-device baselines.
%U https://aclanthology.org/2026.emnlp-main.300/
%P 6904-6919
Markdown (Informal)
[Need to Know: Contextual-Integrity-Grounded Query Rewriting for Privacy-Conscious LLM Delegation](https://aclanthology.org/2026.emnlp-main.300/) (Huang et al., EMNLP 2026)
ACL