PrivacyGuard: Source-Linked Sanitization and Locally Resolvable PII References for Cloud GUI Agents
New in LLM agents and tool use; no direction yet. LLM agents and tool use
Abstract (arXiv)
Cloud graphical user interface (GUI) agents send instructions and screenshots to remote planners, exposing personally identifiable information (PII). Existing placeholder defenses mask detected spans and reuse exact strings. Fragmented or changed forms can expose PII or break entity continuity across turns. PrivacyGuard addresses this private reference problem with source-linked sanitization (SAN) and a structured reference lifecycle (REF). SAN reconstructs fragmented or overlapping evidence and rewrites the complete source extent. REF links changed appearances to session entities and supports locally resolvable references without returning raw values to the upload path. In five-turn sessions scored against common gold, PrivacyGuard reduces audited source-value leakage by 33.55%--82.60% relative to the closest prior full-pipeline system in all four reader--language combinations. REF raises strict cross-form continuity from 0.04 under exact-value reuse to 0.81. In reference-required Android tasks, verified completion rises from 8/30 and 7/30 to 17/30 and 20/30. On AndroidLab, PrivacyGuard achieves the highest or tied-highest task-success point estimate among protection pipelines for all four planners. On PrivScreen, it reduces fixed-reader field exposure by 62.72%--62.84% relative to the same baseline. PrivacyGuard lowers upload exposure while preserving a locally resolvable task path.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 33 of its 37 references that are in the map. First posted on arXiv 12 Oct 2026.
Builds on
33 of 37 references.
References · 33
- 2025UI-TARS: Pioneering Automated GUI Interaction with Native Agentscited by 610
- 2024SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agentscited by 567
- 2023AppAgent: Multimodal Agents as Smartphone Userscited by 415
- 2024Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perceptioncited by 293
- 2024OmniParser for Pure Vision Based GUI Agentcited by 198
- 2025Mobile-Agent-v3: Fundamental Agents for GUI Automationcited by 190
- 2024Mobile-Agent-v2: Mobile Device Operation Assistant with Effective Navigation via Multi-Agent Collaborationcited by 189
- 2024Large Language Model-Brained GUI Agents: A Surveycited by 162
- 2024AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agentscited by 107
- 2024ScreenAI: A Vision-Language Model for UI and Infographics Understandingcited by 98
- 2023GLiNER: Generalist Model for Named Entity Recognition using Bidirectional Transformercited by 88
- 2023Flocks of Stochastic Parrots: Differentially Private Prompt Learning for Large Language Modelscited by 83
21 more
- 2026Mobile-Agent-v3.5: Multi-platform Fundamental GUI Agentscited by 79
- 2024AppAgent v2: Advanced Agent for Flexible Mobile Interactionscited by 68
- 2025The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injectionscited by 59
- 2025AppAgentX: Evolving GUI Agents as Proficient Smartphone Userscited by 46
- 2023Hide and Seek (HaS): A Lightweight Framework for Prompt Privacy Protectioncited by 38
- 2025Towards Trustworthy GUI Agents: A Surveycited by 34
- 2023Split-and-Denoise: Protect large language model inference with local differential privacycited by 32
- 2024PrivacyRestore: Privacy-Preserving Inference in Large Language Models via Privacy Removal and Restorationcited by 28
- 2025PrivWeb: Unobtrusive and Content-aware Privacy Protection For Web Agentscited by 9
- 2025DualTAP: A Dual-Task Adversarial Protector for Mobile MLLM Agentscited by 9
- 2023InferDPT: Privacy-Preserving Inference for Closed-box Large Language Modelcited by 8
- 2026Anonymization-Enhanced Privacy Protection for Mobile GUI Agents: Available but Invisiblecited by 6
- 2026PIIBench: A Unified Multi-Source Benchmark Corpus for Personally Identifiable Information Detectioncited by 6
- 2025CORE: Reducing UI Exposure in Mobile Agents via Collaboration Between Cloud and Local LLMscited by 5
- 2024DiffSTR: Controlled Diffusion Models for Scene Text Removalcited by 3
- 2026WebPII: Benchmarking Visual PII Detection for Computer-Use Agentscited by 3
- 2026A System for Name and Address Parsing with Large Language Modelscited by 2
- 2024Improvement in Semantic Address Matching using Natural Language Processingcited by 1
- 2026LLM-Redactor: An Empirical Evaluation of Eight Techniques for Privacy-Preserving LLM Requestscited by 1
- 2026CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agentscited by 1
- 2026SurrogateShield: Beyond Redaction for High-Utility, Privacy-Preserving LLM Interactionscited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents
- 2026SlotGuard: Stop Oversharing Private Local Context in LLM Agent Transcri
- 2026Chain-of-Sanitized-Thoughts: Plugging PII Leakage in CoT of Large Reasoning Models
- 2026MaskClaw: Edge-Side Personalized Privacy Arbitration for GUI Agents with Behavior-Driven Skill Evolution
- 2025PATCH: Mitigating PII Leakage in Language Models with Privacy-Aware Targeted Circuit PatcHing
- 2026ToolMinimize: Auditing and Rewriting LLM Agent Tool Calls to Minimize Privacy Exposure
- 2026AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents
- 2026PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of Visibility