AI governance under threat managing data poisoning, drift, and integrity challenges

Abstract

Artificial Intelligence (AI) is increasingly embedded within supply chain and information systems, where decision-making depends on the quality, integrity and reliability of operational data. While previous studies have examined data poisoning, data drift and data integrity independently, limited research has integrated these governance risks within a single AI governance perspective. This review addresses that gap by synthesising current literature and developing a case-informed governance framework for managing interconnected AI risks in operational environments. The study adopts a structured narrative review supported by published literature and an illustrative supply chain case scenario to examine how data poisoning, data drift and data integrity failures affect AI performance, organisational resilience and governance. Rather than reporting experimental findings, the paper synthesises existing evidence and demonstrates how these risks interact throughout the AI lifecycle. The review demonstrates that effective AI governance extends beyond technical safeguards alone. Maintaining reliable and resilient AI systems requires continuous lifecycle assurance, robust data governance, organisational accountability, and appropriate regulatory oversight. Drawing on these findings, the study proposes an Integrated AI Governance Framework that brings together technical assurance, organisational governance, and internationally recognised governance principles within a unified lifecycle model. This study contributes to AI governance literature by providing an integrated governance perspective that connects technical AI risks with organisational governance and practical implementation considerations. The framework offers guidance for organisations deploying AI within complex data-driven environments.