AI-Resistant Data Theft Prevention for Consumer Electronics via Context Aware Obfuscation and Post-Quantum Blind Signature

The deep integration of Artificial Intelligence (AI) into Consumer Electronic (CE) systems has greatly enhanced service intelligence, yet poses substantial risks of AI-driven data theft. Existing privacy-preserving approaches often face the dilemma of balancing security, computational efficiency, and compatibility with resource constrained CE devices, failing to effectively combat adaptive AI-driven data theft attacks. To tackle this critical challenge, we propose an AI-resistant data theft prevention scheme for CE systems, which innovatively integrates context aware code obfuscation (CACO) and revised lattice-based post-quantum blind signature to construct a synergistic defense mechanism. Specifically, we design a tailored obfuscation mechanism for AI modules in CE devices by fusing dynamic opaque predicates, context aware control flow flattening, and semantic encryption. This mechanism undermines AI-powered reverse engineering and adversarial sample generation, thereby blocking core attack paths of AI-driven data theft. We also incorporate a lattice-based blind signature scheme to leverage the post-quantum security of lattice cryptography, which safeguards data privacy during transmission authentication and robustly defends against AI-driven interception and correlation attacks. Extensive comparative experiments on benchmark CE platforms demonstrate that the proposed scheme achieves the highest resistance rate against state-of-the-art AI-driven data theft attacks, with its computational cost and communication delay reduced by more than 20% compared with existing hybrid solutions. This work establishes a unified end-to-end AI-resistant privacy protection framework and offers a new benchmark for balancing security, efficiency, and compatibility in CE systems.