Abstract
Some of the potential advancements in information and communication technology (ICT), like artificial intelligence (AI) with federated learning (FL), which are particularly used in the transformation of industrial environments, are driven by digital technologies and can increase the efficiency and dependability of systems. The world has recently become more interconnected as a result of digital technology adoption but with some significant drawbacks. One of the major problems that arises as a malevolent possibility is cybercrime, which poses a threat to governments, corporations, and societies-most significantly, civil society. It is now feasible to work with more than two people to integrate Advanced Digital Technology (ADT) and move the foundation in the direction of a strong hierarchy, but in a complicated way, which has turned into a playground for cybercriminals. The idea of zero trust offers some sweetness in demand for threat detection by autonomously based FL, ML, and DL to address such difficult elements. This paper's main objective, however, is to offer an investigative report based on cyber vulnerability detection using AI, ML, and DL in order to shield the ecosystem from malevolent attacks before they happen. Examining machine learning-enabled classifiers and ensembles for network intrusions and autonomous malicious detection is the secondary objective. Therefore, it explains how we can combine the two models to analyze the context of attacks more effectively and efficiently and how to respond by implementing network security and Internet of Things (IoT) security strategies, such as the stop-and-listen method. When developing and implementing an integrated AI-enabled zero-trust intrusion detection architecture into the present industrial IoT ecosystem, there is a significant discussion about difficult issues and potential solutions. Given the notable outcomes of the simulations we obtained, it is evident at the conclusion of this research that the proposed solution is a strong contender for implementation in the real-time industrial setting.