Securing the AI Supply Chain: Using Blockchain For Verifiable AI Model Provenance on Government Clouds

Abstract

The growing adoption of artificial intelligence (AI) in government cloud systems has already intensified worries around the
integrity, authenticity, and security of AI supply chains. Poisoning of data, model manipulation, and the placement of fake
models are malicious interventions that endanger not only the quality of the results produced by AI but also the credibility
of the decision-making process at the governmental level. The article discusses blockchain as a demonstrable system of AI
model provenance, including its ability to offer unalterable records, decentralized trust, and traceability throughout the
AI development lifecycle. Through provenance implemented via blockchain in government cloud platforms, interested
parties can gain visibility of how models are sourced, spot malicious changes, and satisfy compliance mandates without
undermining the scalability or performance. In addition to outlining the strengths and weaknesses of blockchain in
ensuring the supply chain of AI, the discussion also covers a conceptual implementation framework that can be applied
to governmental regulatory requirements. The results suggest that blockchain-based provenance could be an initial
protection layer to robust and credible AI implementation in sensitive settings associated with the public sector.