In this study, we utilized a dataset consisting of 100 spinal X-ray images to develop an AI-based system that automatically classifies the safety of needle placement during spinal procedures. Each image contains a needle insertion, and a clinical expert manually annotated a red reference line. Based on this reference, the model determines whether the needle has entered the patient's nerve area. This approach enables automated safety classification and verifies the feasibility of core components for a procedure-assisting system.