ANALYSIS OF UNSTRUCTURED DATA PROTECTION ON MOBILE DEVICES: CURRENT CHALLENGES AND TRENDS

Abstract

This article examines the challenges of protecting unstructured data on mobile devices in the context of prolonged sessions, where one-time authentication does not ensure an adequate level of security. A conceptual framework is proposed that combines continuous behavioral authentication, data-centric access policies, and a hybrid cryptographic configuration: symmetric encryption for content and asymmetric encryption for keys.Purpose. To substantiate a model that minimizes the risks of unauthorized access to messages, media, attachments, and cached data within an active session. The focus is on integrating AES with ECC, MFA with behavioral verification, as well as continuous monitoring based on touch dynamics, inertial sensor data, feature fusion, scoring and decision systems, risk-adaptive actions, local and federated learning, and reproducible metrics.Methods. A literature review was conducted using Web of Science and Scopus databases; approaches to continuous authentication and sensor fusion were systematized; permission hygiene and the lifecycle of temporary copies were analyzed; reinforcement learning (RL)-based policies for step-up authentication and re-verification were examined; the proposed model was aligned with modern session management standards and cryptographic resilience requirements.Results. One-time authentication does not guarantee the continuity of user identity. Continuous verification enhanced by sensor fusion reduces the “risk window” without requiring excessive explicit actions. Touch-based features demonstrate high effectiveness, while inertial sensors provide background monitoring. RL-driven policies trigger interventions only under elevated risk conditions (reauthentication, preview masking, export suspension). Data-centric access policies reduce the attack surface by accounting for permissions, side-channel sensor leaks, forensic artifacts, and platform vulnerabilities. The regulatory block consolidates the hybrid model “AES for content ↔ ECC for keys/signatures” and complements it with MFA and post-login monitoring.Conclusions. The proposed configuration incorporates three key components: continuous behavioral verification; data-centric access policies (management of previews, caches, temporary copies, permissions); and hybrid cryptography, where symmetric encryption protects content while asymmetric encryption secures keys and digital signatures.The integration of these components ensures a balance between accuracy, latency, and energy efficiency, making the model suitable for corporate environments and secure messengers.