Explainable and Compliant AI for 6G Communication Systems: Auditable Federated GANs with Blockchain-Backed Traceability and Smart Contract Governance

Abstract

The rapid evolution of $\mathbf{6 G}$ communication systems demands artificial intelligence (AI) solutions that are not only adaptive and explainable but also secure and compliant with international standards. A major challenge lies in achieving transparency and regulatory compliance in federated wireless environments while ensuring data privacy and performance. This research introduces a novel architecture that integrates auditable Federated Generative Adversarial Networks (GANs) with explainability frameworks such as SHAP and LIME, supported by a blockchain layer for immutable audit trails and traceable AI decisions across distributed nodes. Smart contracts are employed to enable dynamic policy enforcement and fine-grained access control. The framework is aligned with 3 GPP, ITU-T, and IEEE standards, making it suitable for compliance-driven deployments. Experimental evaluation demonstrates that the proposed architecture enhances model interpretability by up to $81 \%$, reduces security risks through blockchain-based auditing, and maintains competitive performance in key 6 G use cases such as intelligent beamforming, dynamic spectrum allocation, and edge device authentication. The findings confirm that compliant, explainable, and secure AI can be achieved in next-generation wireless networks without compromising efficiency or user trust.