AI & ML
GapLeak: When Privacy-Preserving LLM-based Web Agents Become a Privacy Risk
arXiv cs.CR0 views
arXiv:2611.22223v1 Announce Type: new Abstract: Large Language Model (LLM)-based web agents rely on data minimization and least privilege to protect user privacy. We show that these principles are necessary but insufficient: they constrain authorized sources, but not higher-sensitivity sources that remain reachable. A website can request a field unavailable from authorized sources, creating an information gap. To complete the task, a goal-driven agent may escalate to reachable but unauthorized sources and disclose accumulated sensitive information. We formalize this vulnerability as GapLeak, where an adversary exploits a legitimate web form without prompt injection or other malicious instructions. We evaluate GapLeak on three web-agent frameworks and six LLMs from Anthropic, Google, and OpenAI across recruitment, e-commerce, travel, and healthcare. GapLeak induces an average unauthorized disclosure of approximately 37%. The information gap is the primary driver: eliminating it reduces disclosure by more than 50%. Existing instruction-level defenses based on least privilege partially mitigate the vulnerability but often reduce task success. We propose submission-time verification, which filters unauthorized content before form submission, eliminating over 96% of unauthorized disclosures while preserving task success.
Source:https://arxiv.org/abs/2611.22223PreviousIs 4-Bit Quantization More Than Noise to a Chat Model's Safety Behaviour? Attributing Behavioural Shifts with Matched PerturbationsNextFrom Token Changes to Task Failures: Bit Flips in a Shared KV-Cache Block
Related posts
Toward Planning Representations That Separate Task Structure from Visual Content
0arXiv cs.RO
Spectral Steering: Local Inverse Geometry for Frozen Diffusion Generators
0arXiv cs.RO
Representing Differential Inclusions as Monads: A Framework for Compositional Controller Synthesis
0arXiv cs.RO
AcousticBEV: Acoustic BEV Object Localization for Intelligent Vehicles
0arXiv cs.RO
Reliable Conditioned Hamilton-Jacobi Reachability via Subsolution Learning on Exact Local Windows
0arXiv cs.RO
Dependable Voice Command for Multi-UAV Swarms Using Small Language Models
0arXiv cs.RO