The smart factory environment has been transformed into an Industrial Internet of Things (IIoT) environment, which is an interconnected and open approach. This has made smart manufacturing plants vulnerable to cyberattacks that can directly lead to physical damage. Most cyberattacks targeting smart factories are carried out using malware. Thus, a solution that efficiently detects malware by monitoring and analyzing network traffic for malware attacks in smart factory IIoT environments is critical. However, achieving accurate real-time malware detection in such environments is difficult. To solve this problem, this study proposes an edge computing-based malware detection system that efficiently detects various cyberattacks (malware) by distributing vast amounts of smart factory IIoT traffic information to edge servers for deep learning processing. The proposed malware detection system consists of three layers (edge device, edge, and cloud layers) and utilizes four meaningful functions (model training and testing, model deployment, model inference, and training data transmission) for edge-based deep learning. In experiments conducted on the Malimg dataset, the proposed malware detection system incorporating a convolutional neural network with image visualization technology achieved an overall classification accuracy of 98.93%, precision of 98.93%, recall of 98.93%, and F1-score of 98.92%.
Installation-time integrity assurance is widely used as a foundational trust mechanism in over-the-air (OTA) software update ecosystems. By verifying cryptographic authenticity and artifact immutability at deployment, such mechanisms establish a provable baseline of artifact integrity. However, in physical AI systems operating in dynamic and context-dependent environments, it is not sufficient to assume that deployment-time verification alone guarantees semantic correctness during execution. This paper demonstrates that deployment-bound sufficiency does not necessarily hold under dynamic runtime conditions, particularly in systems where behavior emerges from interaction with dynamic physical environments. We establish a structural separation, showing that satisfaction of installation-time integrity does not necessarily imply execution-time behavioral correctness. Even when deployed artifacts remain authentic and unmodified, contextual interaction and runtime control dynamics may induce behavior that violates intended safety semantics. To address this structural limitation, we introduce the Lifecycle Integrity and Accountability Coupling (LIAC) framework. LIAC reframes integrity assurance as a lifecycle-aligned accountability architecture rather than a static deployment predicate. By preserving lightweight, stage-aligned evidence across creation, distribution, installation, execution, and governance phases, LIAC enables structured post-incident localization of responsibility without requiring continuous runtime monitoring. Together, these results provide a formal foundation for understanding the limits of installation-centric integrity models and present a scalable accountability framework for OTA-enabled physical AI systems.