Understanding the Practices, Perceptions, and (Dis)Trust of Generative AI among Instructors: A Mixed-methods Study in the U.S. Higher Education
Project Overview
The document explores the complexities surrounding the use of generative AI (GenAI) in higher education, emphasizing the diverse perceptions and levels of trust among instructors. While many educators recognize the potential benefits of GenAI—such as personalizing learning experiences, enhancing critical thinking, and improving lesson planning—there is a notable gap between their familiarity with the technology and its actual implementation in instructional tasks. Concerns regarding academic integrity, social justice, and environmental impacts lead to a cautious adoption of GenAI tools. The findings indicate a pressing need for tailored training and institutional support to foster trust among educators and facilitate effective integration of GenAI into curricula. Additionally, significant challenges persist, including issues related to over-reliance on AI, copyright ethics, and the risk of misinformation, underscoring the importance of developing clear policies and guidelines for responsible AI usage in education. Overall, while GenAI holds promise for enhancing educational practices, careful consideration of its implications is essential for maximizing benefits while mitigating potential drawbacks.
Key Applications
Generative AI for academic writing and personalized learning
Context: Higher education settings, including universities and medical training programs, where both students and faculty engage with AI tools for writing assistance and personalized learning experiences.
Implementation: Integration of generative AI tools (like ChatGPT) to assist in academic writing, assess student responses, and provide personalized learning resources through tailored exercises. This includes qualitative analysis of instructor practices and perceptions regarding AI usage, as well as evaluating the accuracy of AI-generated content in medical contexts.
Outcomes: Enhanced writing support and personalized learning pathways, leading to increased engagement and awareness of AI's potential and limitations in educational practices. Improved understanding of AI capabilities in various educational domains.
Challenges: Concerns about academic integrity, risks of misinformation, potential for plagiarism, limited actual usage of AI tools, reliance on AI for writing tasks, and the need for faculty training and integration of AI tools into curricula.
Implementation Barriers
Cultural/Perceptual
Divergent perceptions of trust and distrust in GenAI among instructors, leading to polarized views on its usefulness. Additionally, concerns about the ethical implications of AI, including plagiarism and the impact on academic integrity.
Proposed Solutions: Calibrated trust strategies focusing on education and practical applications of GenAI, and development of clear policies regarding AI use in academic settings.
Institutional Support
Insufficient institutional guidelines and training related to GenAI integration in educational practices, along with a lack of faculty training in using AI tools effectively in their teaching.
Proposed Solutions: Implement comprehensive training programs, establish clear policies for GenAI usage, and provide professional development programs for educators on AI integration in teaching.
Technical
Issues related to the reliability and accuracy of AI-generated content, leading to misinformation.
Proposed Solutions: Implementing rigorous evaluation standards for AI-generated information.
Environmental
Concerns regarding the environmental impact of training and operating AI systems.
Proposed Solutions: Adopting sustainable practices and assessing the ecological footprint of AI technologies.
Project Team
Wenhan Lyu
Researcher
Shuang Zhang
Researcher
Tingting
Researcher
Chung
Researcher
Yifan Sun
Researcher
Yixuan Zhang
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Wenhan Lyu, Shuang Zhang, Tingting, Chung, Yifan Sun, Yixuan Zhang
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