Language AI in Programming: A Case Study of ChatGPT in Higher Eduation Using Natural Language Processing

Abstract

ChatGPT is an emerging technology used in education. It provides promising support to enhance teaching and learning activities. However, there is lacking understanding of its use to programming courses in computing programs. This study investigates the benefits, issues, and challenges of using ChatGPT, a language AI model, in programming tasks. Data was collected through surveys administered to undergraduate students in computing programs across various National Capital Region (NCR) universities. The collected data was analyzed using Natural Language Processing (NLP) and Latent Dirichlet Allocation (LDA) techniques to gain insights from the students responses. A word intrusion test assessed theme coherence, and evaluators demonstrated moderate agreement (Fleiss Kappa score: 0.62). The study reveals the benefits of employing ChatGPT for programming, including efficient coding, understanding complex codes, and its capability to be used as a problem-solving tool. However, it highlights critical issues and challenges, such as data privacy and ethical concerns, plagiarism tendencies, and contextual understanding limitations. The findings contribute valuable insights for developers, educators, and researchers interested in leveraging language AI models like ChatGPT to enhance programming workflows while effectively addressing associated challenges.

Ähnliche Arbeiten