Understanding Human-AI Trust in Education
Project Overview
The document explores the integration of generative AI, particularly AI chatbots, in education and examines how students develop trust in these systems. It differentiates between human-like trust, rooted in interpersonal relationships, and system-like trust, which relies on the reliability of technology. The research reveals that both trust types significantly shape student perceptions, influencing their engagement, enjoyment, and perceived usefulness of AI chatbots. Notably, human-like trust plays a more critical role in fostering trusting intentions, while system-like trust is more pivotal for behavioral intentions and perceived usefulness. These insights underscore the importance of developing new frameworks to better understand the dynamics of human-AI trust in educational settings, highlighting the potential for generative AI to enhance learning experiences when effectively integrated and trusted by students.
Key Applications
AI chatbots as virtual teaching assistants for coding education
Context: Used in educational settings to provide personalized tutoring, feedback, and code generation. These chatbots assist in explanations of programming concepts and suggestions for code quality, acting as virtual pair programmers.
Implementation: Implemented through large language models that engage in human-like conversations, providing real-time feedback and assistance during coding tasks.
Outcomes: Increased student engagement, improved coding skills, and enhanced understanding of programming concepts.
Challenges: Concerns regarding students' trust calibration, leading to potential over-reliance on AI-generated content or distrust in its accuracy.
Implementation Barriers
Trust calibration
Students may over-trust or under-trust AI systems based on their anthropomorphic characteristics.
Proposed Solutions: Develop educational resources about AI limitations and promote critical evaluation of AI outputs.
Technology reliability
Concerns about the reliability and integrity of AI systems can hinder student engagement.
Proposed Solutions: Improve system consistency and provide clearer information about AI capabilities and limitations.
Project Team
Griffin Pitts
Researcher
Sanaz Motamedi
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Griffin Pitts, Sanaz Motamedi
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