How Large Language Models Are Changing MOOC Essay Answers: A Comparison of Pre- and Post-LLM Responses
Project Overview
The emergence of generative AI (GAI) tools like ChatGPT has notably influenced online education, especially within massive open online courses (MOOCs). A study focusing on student essay responses in a MOOC about AI ethics revealed that the introduction of GAI led to longer essays with reduced vocabulary diversity, although the central themes remained unchanged. This shift raises significant concerns about academic integrity, as students may increasingly depend on GAI for essay writing, potentially undermining the learning process. The findings highlight the necessity for academic institutions to confront these challenges and consider their implications for educational outcomes, suggesting a need for strategies that integrate GAI responsibly while promoting authentic learning experiences.
Key Applications
ChatGPT
Context: Free online course on ethics of AI, targeting university-level students worldwide.
Implementation: Analyzed student essay submissions before and after the release of ChatGPT.
Outcomes: Increased essay length, decreased lexical diversity, and changes in vocabulary usage.
Challenges: Concerns about academic dishonesty and reliance on AI-generated content.
Implementation Barriers
Ethical / Technological
Increased academic dishonesty, reliance on AI tools for completing assignments, and challenges in detecting AI-generated text.
Proposed Solutions: Implement proctored exams, develop better detection mechanisms for AI-generated content, and advance LLM-detection technologies alongside countermeasures in assessment methods.
Project Team
Leo Leppänen
Researcher
Lili Aunimo
Researcher
Arto Hellas
Researcher
Jukka K. Nurminen
Researcher
Linda Mannila
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Leo Leppänen, Lili Aunimo, Arto Hellas, Jukka K. Nurminen, Linda Mannila
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