Plagiarism and AI Assistance Misuse in Web Programming: Unfair Benefits and Characteristics
Project Overview
The document explores the implications of generative AI in education, particularly focusing on its role in web programming. It presents findings from a controlled experiment assessing student performance across three modes: independent work, AI-assisted tasks, and plagiarized submissions. The results reveal that AI assistance and plagiarism resulted in similar test scores to independent efforts, yet significantly reduced the time required to complete assignments, raising serious concerns about academic integrity. Additionally, while students recognize the potential benefits of AI assistance, they remain doubtful about its effectiveness and readability. The study underscores the necessity for automated tools to detect academic misconduct, emphasizing the challenge of maintaining integrity in an educational landscape increasingly influenced by AI technologies.
Key Applications
AI assistance in programming tasks (e.g., ChatGPT)
Context: Web programming education for university students
Implementation: Controlled experiment comparing independent work, AI-assisted work, and plagiarized submissions
Outcomes: Students using AI assistance completed tasks faster but struggled with readability and correctness; they perceived AI as potentially helpful with proper acknowledgment.
Challenges: AI-generated code was often complex and not easily understood; integrating AI solutions into personal work was difficult.
Implementation Barriers
Technical Barrier
Challenges in distinguishing AI-assisted submissions from original work due to their unique nature.
Proposed Solutions: Development of an AI-assistance detector that considers code anomalies and quality metrics.
Educational Barrier
Students' pressure to complete tasks quickly may lead to higher instances of plagiarism and AI misuse. To alleviate this pressure, incentivizing early submissions and breaking assessments into smaller tasks can be beneficial.
Proposed Solutions: Incentivizing early submissions and breaking assessments into smaller tasks to alleviate pressure.
Project Team
Oscar Karnalim
Researcher
Hapnes Toba
Researcher
Meliana Christianti Johan
Researcher
Erico Darmawan Handoyo
Researcher
Yehezkiel David Setiawan
Researcher
Josephine Alvina Luwia
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Oscar Karnalim, Hapnes Toba, Meliana Christianti Johan, Erico Darmawan Handoyo, Yehezkiel David Setiawan, Josephine Alvina Luwia
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