A statistical analysis of engineering students’ usage and perceptions of chatgpt in engineering education

This study examines how engineering students use ChatGPT for academic work and how they evaluate its benefits, risks, and place within engineering education. The manuscript is based on a structured survey of 200 engineering students, predominantly from electrical, electronics, communications, and computer-related fields. The uploaded draft reports near-universal academic adoption, with 98.8% of respondents indicating ChatGPT use for academic purposes and 96.5% reporting more than six months of use. Descriptive results show that students most frequently use ChatGPT for understanding theoretical concepts, explaining formulas and algorithms, writing or debugging code, and preparing for exams. Perception measures indicate a strongly positive overall learning impact (M = 4.16/5), alongside high agreement for asking follow-up questions (M = 4.35), brainstorming support (M = 4.15), improved confidence (M = 4.14), and problem decomposition (M = 4.02). The perception scale demonstrated good internal consistency (Cronbach’s α = 0.835). The draft’s inferential summary also reports a statistically significant positive impact relative to the neutral midpoint (t = 9.6, p < 0.001). Qualitative responses indicate that students value step-by-step explanations, conceptual simplification, coding assistance, and time savings, while continuing to express concerns about hallucinated information, shallow understanding, and dependency. Overall, the findings support guided not unrestricted integration of generative AI into engineering learning. The paper argues that engineering programs should pair AI access with verification practices, prompt literacy, and assessment designs that reward reasoning, transparency, and original problem solving.