Researchers at Korea's Toyo Mirai University have published a new paper on arXiv exploring the ChatGPT educational impact in the field of computer science. The results of the study support the idea thatGenerative AI(gen AI) can help students improve their learning, although some fear the technology could be misused by cheaters.
In this study, researchers Hee-hwan Son, Yudu Kim, and Hee-jin Lee selected 36 computer science majors for a four-week software development competition in which the students could use ChatGPT.
In the first round, all 36 students were evaluated based on performance metrics such as code quality, innovation, and compliance with program requirements. The 15 students with the highest scores advanced to the second round, where the final rankings were made and the winner was selected.
The researchers said they found that at different stages of the project, including ideation, documentation, software development and quality assurance, theStudents who use ChatGPT extensively have higher completion rates and higher scores.. A comparative analysis of the first and second round students found significant differences in their experiences with generative AI in software development.
In a post-study survey, researchers found that the78% participants felt that AI helped them gain a deeper understanding of complex topics. Three-quarters of students reported that the ChatGPT significantly improved their practical skills, and the same percentage reported that it had a positive impact on their professional development.
Generative AI, in its current stage of development, can work closely with developers to work together and produce better results. For example, it can help developers solve problems encountered in the code, speed up the code writing speed, etc., and both sides cooperate with each other to realize complementary advantages.