Crnobrnja, JelenaJelenaCrnobrnjaHorvat, DjerdjDjerdjHorvatRomero, DavidDavidRomeroRuppert, TamásTamásRuppertMarjanovic, UgljesaUgljesaMarjanovic2025-09-082025-09-082026https://publica.fraunhofer.de/handle/publica/49497410.1007/978-3-032-03538-7_6This study examines the adoption of Industry 5.0 technologies and workforce training practices in Serbian manufacturing firms, addressing a critical gap in research on human-centric smart manufacturing in emerging economies. The study aims to answer the key question of how Serbian manufacturing firms adopt Industry 5.0 technologies while developing the necessary workforce skills to support them. Using the Serbian dataset from the 2022 European Manufacturing Survey, we analyze the adoption patterns of six digital/artificial intelligence (AI) tools and five types of training programs across 147 manufacturing firms. Results reveal significant asymmetries in Industry 5.0 technologies implementation. While 50.3% of firms use digital tools for production management, only 10.2% incorporate AI functionalities in this area. Task-specific training is widespread (74.8% adoption), but more strategic training (e.g., innovation-focused programs at 29.9%) remains limited. Most critically, only 6.2% of firms achieve balanced adoption, combining multi-area AI implementation with comprehensive training. The findings demonstrate that Serbian manufacturers predominantly adopt either technologies or training - rarely both - creating potential bottlenecks for effective human-AI collaboration. These results contribute to Industry 5.0 literature by empirically validating implementation challenges specific to emerging economies and highlighting the need for integrated technology-workforce strategies. The study provides practical insights for policymakers and firm managers aiming to accelerate the adoption of Industry 5.0 technologies while maintaining human-centric production systems.enIndustry 5.0Digital TransformationHuman-Centric ManufacturingTechnological AdoptionWorkforce Skills DevelopmentArtificial Intelligenceconference paper