Troubleshooting and fault finding are critical skills for factory workers, as machine downtime contributes to unsafe working conditions and production loss. With the advent of Industry 5.0 (I5.0), traditional training methods become inadequate since there is a pressing need to create human-centric and resilient manufacturing solutions that utilize technology to prioritize human benefits and growth beyond sole system automation while achieving factory objectives like fault-finding. Our study contributes to this need by generating a framework to train I5.0 factory workers in troubleshooting skills using generative AI. The chatbot factory assistant framework integrates the GPT-4 large language model (LLM) for natural language interaction across three functional modules to ensure robustness and data confidentiality. The first module is a GPT-based online chatbot with rich interactive feedback. The second module is a Langchain-based offline chatbot that processes factory data internally for secure responses, and the third one is a backup recovery module that provides summarized equipment health status during service delays. A case study using the AI4I 2020 Predictive Maintenance dataset demonstrates how the system transforms static maintenance records into a dynamic, personalized training experience and highly promotes the practical integration of human-friendly AI tools into factory training processes for I5.0.