Papers for

privacy-focused software teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

MeReader provides offline AI help for private narrative book reading

MeReader: An Offline, Privacy-Preserving, Progress-Aware AI Assistant for Narrative eBook Reading

Abstract: This paper introduces MeReader, a privacy-preserving, offline and progress-aware AI assistant designed to function as a contextual companion for narrative reading on consumer hardware. Unlike conventional Large Language Model (LLM)-based systems that process documents/books online in an all-texts-at-once type of way, thereby risking narrative spoilers, MeReader confines both retrieval and generation to the reader's distinct position in the text. The system integrates a hybrid retrieval architecture, combining vector similarity, BM25, and summary embeddings, to ensure high fidelity to the reader's current context. We evaluate the system across classical English books using both quantitative comprehension metrics and qualitative LLM-based independent judging. The results identify a quality-latency frontier, showing that mid-sized local models (2b-4b) can achieve comprehension parity with larger counterparts (5b+) while maintaining acceptable response quality. Importantly, the proposed progress-aware boundary mechanism is shown to prevent future content leakage without compromising retrieval quality and correctness. These findings suggest that a localized, privacy-centric approach can augment, rather than replace, the digital reading experience, offering a good alternative to cloud-dependent services. Therein, there remains a lack of amalgamate systems which combine progress-aware, privacy-preserving, and narrative eBook reading with AI support. It is in this very gap that the development of MeReader came to be; synthesizing said strands into an offline, book-constrained assistant focused on enhancing comprehension without compromising narrative continuity or user privacy.

Fri 13 NovInformation RetrievalArtificial IntelligenceComputation and Language
The gist
Reading a book with AI help usually means sending your text online, risking spoilers and privacy. The authors present MeReader, an AI assistant working completely offline on your own device, which only looks at the part of the book you're currently reading. This approach keeps your story safe from future spoilers and your data private. They tested MeReader on classic English books, finding that modestly sized local AI models can match bigger ones in understanding while keeping quick response times.
Open → 2611.18513v1

Mobile privacy attitudes shift from acceptance to resignation over time

From Neglect to Resignation: Understanding the Evolution of Mobile Privacy

Abstract: This decade-long longitudinal qualitative study follows 31 smartphone users in China across four interview waves (2014-2023) to examine how mobile privacy understanding evolves alongside shifting computing infrastructures. Findings reveal three stages shaped by distinct interactional mechanisms: early functional acceptance of permission-based data access, growing vigilance triggered by behavioral tracking feedback loops, and eventual resignation as AI-driven inference made ecosystem-level profiling feel inescapable. Rather than a stable attitude, digital resignation emerges through cumulative infrastructure exposure, as each new system layer retroactively reinterpreted earlier permissions and exhausted users' capacity for self-management. The study challenges permission-centric design assumptions and proposes inference visibility, context isolation, and automated privacy agency as system-level responses.

Sat 17 OctHuman-Computer Interaction
The gist
People using smartphones in China have changed how they think about mobile privacy over nearly ten years. At first, they accepted app permission requests without much worry. Then, they became more cautious when they noticed apps tracking them. Eventually, many gave up trying to control their data because advanced AI made it feel impossible to avoid being tracked. The authors show that privacy feelings are shaped by complex systems growing over time rather than just simple permission choices.
Open → 2610.24389v1

LLM agents let users build their own personalized recommender systems

AgentRecommender: LLM Agents Enable Customizable Recommender Systems on the User Side

Abstract: Recommender systems have traditionally been developed for platforms. However, this has given rise to many phenomena that may be advantageous for platform lock-in but are a nuisance to users, such as clickbait, filter bubbles, and the spread of fake news. Recently, user-side recommender systems have been proposed as a new paradigm for solving this problem. If users deploy their own recommender systems, they are no longer at the mercy of the platform's interests. However, building a user-side recommender system is not trivial; in particular, customizing one for oneself requires additional data. We propose AgentRecommender, a method that leverages the investigation capability and internal knowledge of LLM agents to flexibly build user-side recommender systems without additional data. AgentRecommender allows users to easily create recommender systems tailored to their own preferences.

Fri 25 SeptInformation RetrievalArtificial IntelligenceDatabases
The gist
Online recommendation systems usually work for the platforms that run them, which can lead to problems like fake news and biased content. The authors present a new approach where people can run their own recommendation systems on their devices using smart AI agents called LLMs. This method, called AgentRecommender, uses these AI agents' existing knowledge and abilities to tailor recommendations without needing extra data from the user. This way, users are in control and get suggestions that match their personal preferences better.
Open → 2609.31166v1