Privacy-Preserving AI for Detecting and Mitigating Customer Price Discrimination in Big-Data Systems

Abstract

With the rapid development of big data and artificial intelligence technologies, personalized pricing has become a common strategy for platforms to enhance profitability. However, this practice often evolves into customer price discrimination (CPD), which seriously harms consumer rights and market fairness. Meanwhile, the data-driven nature of AI models for CPD detection raises severe privacy concerns. To address the dual challenges of CPD governance and data privacy protection, this paper proposes a privacy-preserving AI framework for CPD detection and mitigation in big-data systems. First, we design a federated learning-based detection model that enables multiple data holders to collaborate on model training without sharing raw user data. Second, differential privacy technology is integrated into the model training process to prevent sensitive information leakage from gradient calculations. For CPD mitigation, a dynamic pricing calibration mechanism based on explainable AI is proposed to ensure pricing transparency while maintaining platform operational efficiency. Experimental results on real-world e-commerce and ride-hailing datasets show that the proposed framework achieves a detection accuracy of 87.6% for CPD behaviors, with a privacy budget consumption of only 1.2, which outperforms traditional centralized models and privacy-unaware AI models. This research provides a technical solution for balancing personalized services, market fairness, and data privacy protection in big-data systems. Notably, the framework exhibits strong cross-platform adaptability, achieving detection accuracies of over 85% on both e-commerce and ride-hailing scenarios with minimal parameter adjustments, making it suitable for deployment in diverse service industries such as online travel and digital content platforms. In practical applications, the framework helps platforms comply with strict data protection regulations including the EU's GDPR and China's Personal Information Protection Law, while reducing consumer complaints related to price discrimination by an estimated 42%. Additionally, the integration of blockchain-based traceability ensures that pricing adjustment records are tamper-proof, providing regulatory authorities with credible audit trails for market supervision.