Johan Stahre

The increasing integration of artificial intelligence (AI) in manufacturing is reshaping human roles on the shop floor. While AI has the potential to enhance human capabilities, questions arise about its impact on workers. Ensuring human-centricity remains crucial, but there is a need to investigate how these principles translate to human-AI collaboration. This paper presents a structured literature review of 25 papers, examining human-centric goals, challenges and enablers in human-AI collaboration. The results show that the most emphasized goals include trust, transparency, augmentation, explainability, safety, and well-being. Prominent challenges include job displacement, conflicting priorities, and algorithmic bias. Identified enablers include co-design approaches, explainable interfaces, and dynamic task allocation. The findings highlight that human-centric aspects are interconnected, requiring a holistic approach. Although human-centricity is widely advocated, it is often framed at the system interface level, rather than in organizational or design processes. Future research should deepen the understanding of cognitive and psychosocial dimensions of human-AI collaboration and develop more empirical and longitudinal insights.

Immersive technologies are entering operator training on the assumption that a more engaging modality produces better learning. Whether the modality itself changes what learners keep has rarely been tested against simpler or AI-agent alternatives. We compared three ways of delivering the same 16-step assembly procedure (immersive virtual training, a humanoid social robot driven by a large language model, and interactive video) in a between-subjects experiment with 36 novices. Each participant learned the procedure under one modality and then performed the task once without instruction. We recorded perceived subjective effort (RSME), heart-rate variability (HRV) normalised to each person's baseline, completion time, and error rates. During learning, perceived subjective mental effort did not differ between media, and VR was rated the easiest to follow. The central result is a dissociation between how learning felt and how it turned out: the instruction modality users rated easiest, VR, produced the slowest and least accurate performance once support was removed, and effort during learning was unrelated to later performance (ρ = -0.02). The mental workload imposed by an instruction modality is therefore not revealed by how effortless it feels subjectively during learning phase but a misalignment pattern. Its distinctive contribution, however, is a human-centric assessment of emerging digital technologies, together with a comprehensive, multi-measure evaluation approach intended to inform future industrial practice to adopt new learning technologies for Industry 5.0.

1. With the current global wars, Sweden’s accession to NATO, and defense spending to rise by 300 million Swedish kronor in the next decade (3.5% of total GDP), the conditions to strengthen Swedish ...

Effective human–AI interaction in manufacturing requires that AI systems do more than deliver information—they must adapt to the operator's cognitive state, augmenting rather than overloading human performance. This study presents the first comparative empirical investigation of mental workload across three distinct human–AI interaction modalities—Virtual Reality (VR), an AI-integrated humanoid social robot (Furhat + GPT-4o), and video instruction—evaluated during both learning and operational phases of an industrial assembly task (n = 36). Using a triangulated measurement protocol combining subjective ratings (RSME), physiological monitoring (heart rate variability via PPG), and performance outcomes (completion time, error rate), we identify a critical dissociation between perceived and physiological cognitive load: VR produced the lowest perceived effort during learning (RSME median = 25.50) yet the worst operational performance (completion time median = 490 s; error rate median = 8), while video yielded consistently low workload and the highest accuracy. The humanoid robot occupied a distinct middle position, generating the highest HRV ratios during learning (median = 1.02)—physiological evidence of social engagement not captured by self-report—alongside moderate performance outcomes comparable to video. Crucially, VR's low perceived effort masked a hidden physiological cost that subjective measures entirely failed to detect, demonstrating that self-report alone is insufficient to evaluate human–AI interaction quality. The findings surface a fundamental design principle for human–AI symbiosis: different interaction modalities engage distinct cognitive and autonomic processes, and only triangulated measurement can reveal this. Together, the results provide empirically grounded guidance for designing adaptive human–AI training systems in Industry 5.0 manufacturing.

Since the introduction of Industry 4.0, digital twin technology has advanced dramatically, and its applications have been extensively discussed. Through digital twins, real-time connected production systems are anticipated to be more efficient, resilient, and sustainable, facilitating communication and connectivity between digital and physical systems. However, challenges remain in environmental performance and in integration of virtual reality (VR) and artificial intelligence (AI). Further exploration of digital twin technologies is needed to validate the real-world impact and benefits. This paper investigates these challenges by implementing a real-time digital twin based on the ISO 23247, Automation Systems and Integration - Digital Twin Framework for Manufacturing, (Digital twin framework for manufacturing), connecting the physical factory and simulation software with VR capabilities. This digital twin system provides cognitive assistance and a user-friendly interface for operators, thereby improving cognitive ergonomics. The connection of the Internet of Things platform allows the digital twin for real-time bidirectional communication, collaboration, monitoring, and assistance. A laboratory-scale drone factory was used as the digital twin application to test and evaluate the ISO 23247 standard as well as its potential benefits. Additionally, AI integration and environmental performance key performance indicators have been considered as the next stages to improve VR-integrated digital twins. This paper further discusses how digital twins contribute to the improvement of the environmental performance of production systems, primarily from the perspective of data inventory management. Supported by a solid theoretical foundation and a laboratory-scale demonstration of the VR-integrated digital twins in a drone framework, the paper bridges the gap between possible technological integrations and the advancement of digital twins framework based on ISO 23247.

The transition to Industry 5.0 requires a paradigm shift from technology-driven production to human-centric, sustainable, and resilient systems. Although the theoretical foundation of Industry 5.0 is strong, empirical evidence on the workforce readiness required for this transition remains scarce. This study addresses this gap by developing an empirical typology of workforce readiness, based on structured interviews with 160 European industrial stakeholders, using the European Commission’s Industry 5.0 Learning and Assessment tool. Through cluster analysis, we identify three distinct organisational profiles: Ad hoc, Structured, and Adaptive. These clusters reveal significant diversity in how firms approach skill development, the skill levels of their workforces, and their workforce resilience. The findings show that while Structured firms have established upskilling programmes, they still lack in-demand skills such as AI literacy, whereas Ad hoc firms struggle with foundational knowledge management. The typology provides a framework for managers to develop their human-centric transition strategies and for policymakers to design targeted incentives, moving beyond generic upskilling policies.

As industry navigates the transition towards Industry 5.0, the need for human-centric workforce development becomes more pressing than ever. This shift highlights not only technological implementations but also employee empowerment through lifelong learning. Digital platforms for learning and skill management rapidly emerge, to align individuals’ skills with changing organisational needs. However, how these platforms are perceived and integrated across different companies is greatly underexplored. This study describes a mixed-methods evaluation of a digital skill management platform implemented in two industrial contexts: large Swedish companies and small and medium-sized enterprises (SMEs) in Germany. The platform offers individual skill gap assessments, recommending suitable training for users. In the study, two main dimensions were investigated, i.e. the platform’s perceived effectiveness for users and the organisational integration of the platform. Results show a range of challenges in said dimensions. Users report varying platform usability and relevance of recommended learning. Results also reveal challenges in organisational integration and in aligning the platform with existing company-internal skill, task, and job taxonomies, including learning management systems (LMS). Other identified issues were user engagement and the visibility of learning progress for managers. Our conclusions emphasise the importance of adapting solutions to organisational contexts when implementing skill management systems. Based on insights from the study, recommendations for supporting successful implementation of such platforms, including the need to accommodate different stakeholder needs, are outlined. Our recommendations include enabling conditions, organisational integration and platform design. The presented results contribute to ongoing discussions on how digital learning solutions can support efficient workforce transformation towards Industry 5.0. Our findings have clear value for the university and research institution context, where researchers, teachers, and leaders are exploring ways to deliver lifelong learning opportunities in industry. Skill management platforms have potential to serve as bridges between university courses and workplace learning demands, enabling universities to play important roles in their employees’ lifelong learning.