Virtual AI assistants in the manufacturing industry – innovation diffusion perspective

The study provides empirical insights into the innovation diffusion of AI assistants within the manufacturing industry. This study presents an innovative proof-of-concept (PoC) solution for a virtual AI assistant supporting human-cobot assembly tasks in manufacturing. Large Language Models (LLMs) are utilized for delivering operator instructions and Vision-Language Models (VLMs) support safety monitoring. The innovation diffusion analysis of the PoC solution identified several benefits, including improved assembly efficiency, reduced errors and hesitation, and flexible, hands-free instruction via natural language. Promising future applications of AI assistants in various industrial operations may include e.g. training, product documentation, and the sharing of tacit knowledge. To accelerate the diffusion of such innovations, further exploration is needed into key issues such as AI tool reliability, cybersecurity, compatibility with existing operations, and alignment with user needs.