Harnessing LLMs in Curricular Design: Using GPT-4 to Support Authoring of Learning Objectives
Project Overview
The document explores the application of GPT-4, a generative pre-trained transformer, in the educational context, specifically for creating high-quality learning objectives (LOs) for a university course on Artificial Intelligence. It addresses the challenges educators encounter when developing effective LOs and illustrates how GPT-4 can generate sensible, measurable objectives that align with Bloom's taxonomy, thereby aiding in curricular design. This innovative approach not only alleviates the workload for instructors but also enhances the overall quality of educational planning. However, the authors caution against potential drawbacks, such as the risk of losing pedagogical nuance and the necessity of human validation to ensure the appropriateness of the generated content. Ultimately, the findings suggest that while generative AI has the potential to significantly improve educational outcomes, careful consideration must be given to its integration in teaching practices to maintain the integrity and effectiveness of the learning process.
Key Applications
Using GPT-4 to automatically generate learning objectives
Context: University course on Artificial Intelligence, targeting educators and curricular designers
Implementation: GPT-4 was prompted with guidelines for structuring LOs based on Bloom's taxonomy, producing LOs for conceptual modules and projects.
Outcomes: Generated LOs were mostly sensible, well-expressed, and appropriate to the expected levels of Bloom's taxonomy, improving the course design process.
Challenges: Some generated LOs lacked specificity and the potential for over-reliance on automated systems may stifle creativity.
Implementation Barriers
Technical
The generated learning objectives may lack specificity and clarity, which could lead to misunderstandings or ineffective teaching strategies.
Proposed Solutions: Further tuning of prompts and human validation of generated LOs to ensure relevance and accuracy.
Ethical
Concerns exist regarding the ethical implications of using LLMs trained on copyrighted materials.
Proposed Solutions: Establishing clear policies and guidelines for the ethical use of LLMs in educational contexts.
Pedagogical
There is a risk of losing pedagogical nuance and adaptability if educators overly rely on automated systems, potentially undermining the expertise of educators.
Proposed Solutions: Using LLMs as supportive tools rather than replacements for educators' expertise.
Project Team
Pragnya Sridhar
Researcher
Aidan Doyle
Researcher
Arav Agarwal
Researcher
Christopher Bogart
Researcher
Jaromir Savelka
Researcher
Majd Sakr
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Pragnya Sridhar, Aidan Doyle, Arav Agarwal, Christopher Bogart, Jaromir Savelka, Majd Sakr
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