Artificial Intelligence

The paper specifies the genAI support needs for industrial maintenance against the background of a sociotechnical systems perspective. Emphasizing two needs, accessing implicit operator knowledge and prioritizing complex regulatory knowledge, a multi-layer architecture is outlined for an AI-based context-sensitive maintenance assistance system (MAS). The main purpose is to bridge knowledge gaps with genAI if human expertise and human implicit knowledge are not available and to cope with sub-process-specific challenges of multiple regulations. The MAS facilitates access to technical knowledge, distributes expertise, and shares implicit knowledge of experienced operators across different layers of information processing. The approach goes beyond standardization and has a high potential to enhance organizational as well as individual resilience.

The increasing complexity of industrial environments demands new competencies from workers, particularly the ability to interact with advanced digital systems. Traditional training methods often fall short in supporting the effective transfer of applied knowledge to such contexts, and the effectiveness of this transfer, as measured by performance-based outcomes, remains to be investigated. To address this gap, the present study employed a between-subjects experimental design comparing augmented reality- and paper-based instructions within a realistic production training scenario. The results show that participants who learned with augmented reality completed the production process significantly faster and with fewer errors than those using paper instructions. In addition, learners using augmented reality reported higher usability and experienced lower cognitive load during training. These findings suggest that augmented reality can enhance the transfer of practical skills in industrial ...

The engineer-to-order (ETO) production approach is gaining relevance in response to increasing demand for individualized products and small batch sizes. However, ETO inherently reduces the economies of scale typically achieved in series production, as each order requires tailored engineering and production steps. This loss of efficiency can be mitigated through demand-driven and context-aware information provision throughout the product development process. A recommendation system based on semantic artificial intelligence (AI) and machine learning can support this by i) analyzing historical data and prior knowledge, for example drawings or a bill of materials from previous projects, and ii) making automated suggestions, like reusing existing designs or proposing design alternatives, thus compensating for the aforementioned effects.

Despite its potential, the introduction of artificial intelligence (AI) in industry is often delayed, primarily due to perceived complexity, high costs, and a lack of expertise. This article presents a modular demonstrator reference architecture that provides practical, low-cost access to industrial AI applications. Developed within a design science research approach, it specifically supports experimentation, learning, and gradual integration into existing production processes. The focus is on machine vision, implemented using cost-effective hardware and open-source software. Its applicability is demonstrated in three scenarios: quality control, chip classification, and in-company training. Initial evaluations confirm the technical feasibility, didactic relevance, and transferability to a variety of industrial contexts.

Artificial intelligence is increasingly implemented in companies, but often without clear organizational anchoring. This article evaluates centralized, decentralized, hybrid, and project-based frameworks for the structural integration of artificial intelligence in corporate organizations. A decision table provides guidance for selecting suitable models. In the conclusion, further open research questions are posed.

Despite the commercial availability of simulation tools, using of discrete-event simulation for complex production and logistics systems is becoming increasingly challenging. It requires extensive expertise, high data quality, and considerable time and financial resources. For many years, therefore, there has been high demand for methodological and organizational support for the conduction of simulation studies. This article is based on an analysis of relevant publications and aims to classify previous research on improving the use of simulation. It also raises the question of the need for assistance in applying discrete event simulation and identifies areas for action.

The work system is the key unit of analysis within the discipline of human factors/ergonomics (HFE); it is also considered a fundamental element for value creation within other domains. Its concept is based on sociotechnical systems theory and, within HFE, it conveys a distinctly human-centered perspective. So far, work system models have focused on one or several people working within a defined setting as the only (intelligent) actors within the system. The introduction of generative artificial intelligence (genAI) into work systems, particularly as an intelligent and autonomous actor (agent) with potentially specific social abilities and personality traits, calls for reconceptualization. This article elaborates on the new requirements related to the introduction of genAI and develops a human-centered AI-paired work system model that recognizes the significantly expanded capabilities of AI-enabled collaborative social robots.

Industrial companies in Germany face demographic change and stagnating productivity in an increasingly complex world. Manual assembly remains essential for complex, low-volume products, yet productivity and quality lag due to human variability. This paper introduces a concept and demonstrator for an empathic assembly assistance system that merges a human digital twin and AI-based screwdriver data analytics within a modular architecture. Tightening anomalies are classified, linked to inferred worker states and translated into information and recommendations.

Traditional quality control often falls short in high-mix, low-volume production environments due to variability and complexity. This project introduces an advanced workstation to boost industrial productivity and quality, developed with Axiomatic Design to ensure a clear link between customer needs, functional requirements, and design solutions. Combining polarization cameras, high-resolution imaging, adaptive lighting, and deep learning-based computer vision, the system performs high-accuracy inspection on quantity, quality, and compliance. A digital assistance system offers real-time feedback via an intuitive interface. Validation in a controlled environment confirmed both the system’s practical benefits and its scalability.

Compared to traditional AI systems, generative artificial intelligence (GenAI) introduces user-dependent characteristics that create unique challenges for AI governance in organizations. These challenges are particularly tied to human factors, such as employee attitude, awareness, and skills, which are often neglected by existing governance frameworks. This qualitative case study examines how a manufacturing organization implemented GenAI governance mechanisms to foster the responsible use of this technology. The findings reveal that organizations should adopt a holistic approach, combining structural, procedural, and relational mechanisms to address employee-related aspects of GenAI governance. As a result, this study contributes to the growing field of GenAI governance and provides practical insights for its responsible use in organizations.