Student Perspectives on Using a Large Language Model (LLM) for an Assignment on Professional Ethics
Project Overview
The document explores the integration of Large Language Models (LLMs) in a computing course centered on professional ethics, where students utilized the LLM as a dialogue partner to dissect ethical dilemmas. This innovative approach significantly enriched their comprehension of ethical theories and bolstered their academic writing skills. Feedback from students indicated a generally positive experience, with noticeable gains in knowledge; however, they also identified limitations, particularly the LLM's shortcomings in delivering nuanced analyses or definitive solutions to ethical questions. Educators emphasized the importance of providing appropriate guidance on using LLMs to prevent students from developing misconceptions about the models' authority and reliability. Overall, the findings suggest that while generative AI can effectively enhance learning outcomes in educational settings, careful implementation and oversight are crucial to maximizing its benefits and addressing its limitations.
Key Applications
Using ChatGPT as a discussion partner for analyzing ethical dilemmas
Context: Graduate course for computing master's students focusing on professional ethics
Implementation: Students used ChatGPT to discuss and reflect on a fictional ethical case while documenting their conclusions.
Outcomes: Increased knowledge and understanding of ethical theories; enhanced critical thinking; improved academic writing.
Challenges: LLM provided unfeasible options, lacked depth in analysis, and some students perceived it as an authoritative source.
Implementation Barriers
Perception Barrier
Some students viewed the LLM as an authoritative source of knowledge, leading to uncritical acceptance of its outputs. This could result in students not questioning the validity of the information provided.
Proposed Solutions: Emphasize the importance of critically analyzing LLM outputs and provide instruction on how to effectively use LLMs to foster a questioning mindset.
Knowledge Barrier
Students with less experience may trust the LLM's incorrect information, leading to misconceptions about its capabilities and limitations.
Proposed Solutions: Integrate more instructional support on LLM usage, including characteristics, limitations, and the need for critical evaluation of outputs.
Project Team
Virginia Grande
Researcher
Natalie Kiesler
Researcher
Maria Andreina Francisco R
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Virginia Grande, Natalie Kiesler, Maria Andreina Francisco R
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