How Do Programming Students Use Generative AI?
Project Overview
The document explores the integration of generative AI tools, particularly ChatGPT, in the education of programming students, revealing both beneficial and detrimental effects on learning. While students utilize these tools to generate coding solutions and retrieve information, the study indicates a troubling trend of over-reliance that hampers their critical thinking and understanding of the material. This reliance raises significant concerns regarding academic integrity and the overall effectiveness of learning outcomes. The findings emphasize the urgent need for educators to adapt curricula to responsibly incorporate AI technologies, fostering an environment where students can benefit from these tools without compromising their educational development. Additionally, it calls for the enhancement of assessment methods to ensure that students genuinely grasp the concepts being taught, thereby promoting a balanced approach to the use of generative AI in educational settings.
Key Applications
ChatGPT as an AI assistant for coding exercises
Context: Programming students in an introductory software development course
Implementation: Students were given monitored access to ChatGPT while solving programming exercises, specifically designed to be challenging for the AI.
Outcomes: Students often used ChatGPT to generate code solutions, which led to a cycle of incorrect submissions and reliance on AI for fixes, impacting their learning process.
Challenges: Students frequently submitted code generated by ChatGPT without comprehending their mistakes, leading to diminished critical thinking and programming skills.
Implementation Barriers
Educational
Students struggle with over-reliance on generative AI, which hampers their ability to think critically and solve problems independently.
Proposed Solutions: Educators should modify curricula to incorporate generative AI tools effectively, emphasizing teaching strategies that promote understanding and critical evaluation.
Technical
Generative AI tools can produce incorrect or misleading outputs, leading students to trust AI responses over their own understanding.
Proposed Solutions: Instructors should provide guidance on the limitations of generative AI and encourage students to validate AI outputs against their own knowledge.
Project Team
Christian Rahe
Researcher
Walid Maalej
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Christian Rahe, Walid Maalej
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai