The Impact of Blockchain Technology on Tax Transparency and Risk Control in Multinational Enterprises

Abstract

This study addresses key challenges in tax governance for multinational enterprises, including data silos, delayed risk identification, and insufficient privacy protection. It proposes a collaborative governance framework that integrates blockchain-based smart contracts and multimodal machine learning, termed the Blockchain-Enhanced Multimodal Risk Assessment Framework. The framework employs a dual-layer architecture that combines blockchain technology with federated feature engineering. At the data processing layer, it utilizes a multi-chain coordination mechanism based on Adaptive Proof of Stake, enhanced by a node reputation-based dynamic evaluation algorithm to significantly improve consensus efficiency and data transmission security. At the risk prediction layer, it integrates temporal dependency mining with heterogeneous graph neural networks, applying a composite embedding method to extract 89 highly discriminative features from an initial 1,423-dimensional feature space.