Developing an AI-Driven NC Programming Assistant: A Productive Failure Approach in CNC Education

The integration of artificial intelligence (AI) into engineering and technology education offers new possibilities for enhancing student learning and feedback delivery. In machining, where accuracy and logic are essential, AI can serve as a powerful support tool. This paper presents the NC Programming Assistant, an application that combines rule-based logic and generative AI to help students write and debug numerical control (NC) programs. The assistant provides real-time, context-specific feedback on errors such as syntax issues and safety violations, enabling students to revise their work and strengthen their programming skills. Grounded in the Productive Failure Model, the assistant is designed not only to correct errors but also to encourage learners to engage in trial, failure, and reflection. The tool was deployed in a machining course, where its impact was evaluated through surveys, usage data, and statistical analysis. Results showed improvements in student confidence, debugging ability, and comprehension. By providing immediate and personalized feedback, the assistant addresses common instructional challenges in technical education and demonstrates how theory-informed AI tools can enhance learning outcomes in skill-based disciplines.