PulseAugur
实时 10:57:16

New paper: On-device AI privacy risks extend beyond local computation

A new research paper argues that on-device AI in operating systems presents significant privacy risks, even when computation is local. The paper introduces a framework for governing these systems, focusing on institutional accountability rather than just where the AI runs. It proposes a taxonomy of privacy risks, architectural controls, and an audit rubric, which are then applied to analyze Apple Intelligence, Android AICore, and Microsoft Recall. AI

影响 Highlights the need for robust privacy governance in OS-integrated AI, potentially influencing future OS design and user data protection standards.

排序理由 The cluster contains an academic paper discussing AI privacy concerns and proposing a framework. [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

报道来源 [1]

  1. arXiv cs.AI TIER_1 English(EN) · Jonghyun Chung, Sanket Badhe ·

    Local Is Not a Sufficient Privacy Boundary: Governing OS-Integrated On-Device AI

    arXiv:2606.10173v1 Announce Type: cross Abstract: As AI systems move into operating systems, privacy no longer turns only on whether a model runs locally. A local assistant may assemble email, calendar entries, files, screenshots, notifications, and app intents; retain embeddings…