Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2604.21308 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908988182888448 |
|---|---|
| author | Fu, Wenjie Qin, Xiaoting Zhang, Jue Lin, Qingwei Wutschitz, Lukas Sim, Robert Rajmohan, Saravan Zhang, Dongmei |
| author_facet | Fu, Wenjie Qin, Xiaoting Zhang, Jue Lin, Qingwei Wutschitz, Lukas Sim, Robert Rajmohan, Saravan Zhang, Dongmei |
| contents | Enterprise LLM agents can dramatically improve workplace productivity, but their core capability, retrieving and using internal context to act on a user's behalf, also creates new risks for sensitive information leakage. We introduce CI-Work, a Contextual Integrity (CI)-grounded benchmark that simulates enterprise workflows across five information-flow directions and evaluates whether agents can convey essential content while withholding sensitive context in dense retrieval settings. Our evaluation of frontier models reveals that privacy failures are prevalent (violation rates range from 15.8%-50.9%, with leakage reaching up to 26.7%) and uncovers a counterintuitive trade-off critical for industrial deployment: higher task utility often correlates with increased privacy violations. Moreover, the massive scale of enterprise data and potential user behavior further amplify this vulnerability. Simply increasing model size or reasoning depth fails to address the problem. We conclude that safeguarding enterprise workflows requires a paradigm shift, moving beyond model-centric scaling toward context-centric architectures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_21308 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CI-Work: Benchmarking Contextual Integrity in Enterprise LLM Agents Fu, Wenjie Qin, Xiaoting Zhang, Jue Lin, Qingwei Wutschitz, Lukas Sim, Robert Rajmohan, Saravan Zhang, Dongmei Cryptography and Security Computation and Language Enterprise LLM agents can dramatically improve workplace productivity, but their core capability, retrieving and using internal context to act on a user's behalf, also creates new risks for sensitive information leakage. We introduce CI-Work, a Contextual Integrity (CI)-grounded benchmark that simulates enterprise workflows across five information-flow directions and evaluates whether agents can convey essential content while withholding sensitive context in dense retrieval settings. Our evaluation of frontier models reveals that privacy failures are prevalent (violation rates range from 15.8%-50.9%, with leakage reaching up to 26.7%) and uncovers a counterintuitive trade-off critical for industrial deployment: higher task utility often correlates with increased privacy violations. Moreover, the massive scale of enterprise data and potential user behavior further amplify this vulnerability. Simply increasing model size or reasoning depth fails to address the problem. We conclude that safeguarding enterprise workflows requires a paradigm shift, moving beyond model-centric scaling toward context-centric architectures. |
| title | CI-Work: Benchmarking Contextual Integrity in Enterprise LLM Agents |
| topic | Cryptography and Security Computation and Language |
| url | https://arxiv.org/abs/2604.21308 |