Student-AI Interaction: A Case Study of CS1 students
Project Overview
The document investigates the role of generative AI, particularly ChatGPT, in the educational context of Computer Science 1 (CS1) programming courses. Through a case study, it analyzes how students interact with AI tools while engaging in programming tasks, revealing both advantages and drawbacks. The findings indicate that while generative AI can offer immediate assistance and enhance understanding of complex concepts, it also risks fostering dependency, with some students leaning on AI for solutions rather than developing their own problem-solving skills. This reliance may undermine their independent learning efforts and self-efficacy in programming. Overall, the study aims to illuminate the implications of generative AI on student learning outcomes, suggesting that while such technologies can be beneficial, careful integration and pedagogical strategies are essential to mitigate potential negative effects on students' educational experiences.
Key Applications
Integration of ChatGPT as a VSCode plugin
Context: Introductory programming course (CS1) at a large public university, targeting novice programmers
Implementation: Students used a custom VSCode plugin that integrated ChatGPT directly into their coding environment to assist with programming tasks.
Outcomes: Participants interacted extensively with Generative AI, with varying success rates in task completion. Some students reported increased self-efficacy when using AI effectively.
Challenges: Many students relied on generative AI for complete solutions without attempting the tasks independently, raising concerns about over-reliance and insufficient development of programming skills.
Implementation Barriers
Over-reliance on AI
Students frequently submitted full task descriptions to AI tools instead of attempting the tasks themselves, leading to potential skills deficits.
Proposed Solutions: Encouraging iterative problem-solving approaches and teaching students to decompose tasks into manageable parts could help mitigate reliance.
Self-efficacy issues
Some students experienced decreased self-efficacy after using generative AI, particularly those who relied heavily on it for tasks.
Proposed Solutions: Scaffolded interactions with AI that promote self-regulation and encourage independent problem-solving could improve self-efficacy.
Project Team
Matin Amoozadeh
Researcher
Daye Nam
Researcher
Daniel Prol
Researcher
Ali Alfageeh
Researcher
James Prather
Researcher
Michael Hilton
Researcher
Sruti Srinivasa Ragavan
Researcher
Mohammad Amin Alipour
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Matin Amoozadeh, Daye Nam, Daniel Prol, Ali Alfageeh, James Prather, Michael Hilton, Sruti Srinivasa Ragavan, Mohammad Amin Alipour
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