Bridging the AI Adoption Gap: Designing an Interactive Pedagogical Agent for Higher Education Instructors
Project Overview
This document explores the integration of generative AI in higher education through the development of an interactive pedagogical agent powered by large language models (LLMs). It highlights the importance of human-centered design to facilitate AI adoption among instructors, especially those with limited AI literacy and skepticism towards technology. The research aims to meet the specific needs of educators by offering customized and contextually relevant teaching recommendations, thereby addressing the existing gap in AI adoption and enhancing instructional practices. Key findings indicate that fostering social transparency, implementing incremental information collection, and validating AI-generated suggestions with pedagogical experts are crucial for successful integration. Ultimately, the study underscores the potential of generative AI to support educators in improving teaching effectiveness while also addressing the challenges and concerns associated with its implementation.
Key Applications
Interactive Pedagogical Agent (Chatbot)
Context: Higher education instructors with varying levels of AI literacy
Implementation: Participatory design sessions with pedagogy experts to develop and evaluate chatbot interactions and LLM-generated teaching suggestions
Outcomes: Enhanced trust among instructors, improved AI adoption rates, and tailored teaching suggestions that align with pedagogical practices.
Challenges: Resistance from AI-conservative instructors, data privacy concerns, and the need for ongoing expert validation of AI-generated content.
Implementation Barriers
Technical
Instructors' varying levels of AI literacy and negative attitudes towards AI hinder adoption.
Proposed Solutions: Provide tailored support and training workshops; design AI tools with incremental engagement to accommodate different proficiency levels.
Institutional
Institutional policies framing AI as a threat rather than an asset, leading to skepticism among instructors.
Proposed Solutions: Shift institutional narratives to promote AI as a pedagogical asset; establish clear guidelines and support for AI integration in teaching.
Social
Concerns about pedagogical autonomy and data privacy among instructors.
Proposed Solutions: Incorporate peer validation and social transparency in AI tools; clarify data handling policies upfront to build trust.
Project Team
Si Chen
Researcher
Reid Metoyer
Researcher
Khiem Le
Researcher
Adam Acunin
Researcher
Izzy Molnar
Researcher
Alex Ambrose
Researcher
James Lang
Researcher
Nitesh Chawla
Researcher
Ronald Metoyer
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Si Chen, Reid Metoyer, Khiem Le, Adam Acunin, Izzy Molnar, Alex Ambrose, James Lang, Nitesh Chawla, Ronald Metoyer
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