Cyber security culture as a resilience-promoting factor for human-centered Machine Learning and Zero-defect Manufacturing environments

Abstract

Humans have often been perceived as a leading cause of error in Zero-defect manufacturing (ZDM) processes. There is thus a reduction of human interventions in the deployment of industry 4.0 (I4.0) technologies used for ZDM such as Machine Learning (ML). However, as manufacturing (e.g., I4.0 context) is often placed within a socio-technological context involving the co-integration of humans and technology, the manufacturing processes are now more vulnerable to cyber risk and threats. System vulnerabilities also derive from limitations associated with ML. This paper highlights three challenges associated with ML: explainability, data privacy, and security for ZDM. We argue that due to the high level of data complexity and lack of flexibility in ML models, humans play a critical role in ZDM decision-making. The paper explores the concept of security culture as an enabler for transformative resilience and zero-defect manufacturing and contributes to rethinking the human-centered approach in ZDM. The paper stresses a need to enhance contextual and empirical understanding of transformative resilience and security culture in ML/ZDM environments to better address adverse events such as cyber threat situations.