Embracing AI in Education: Understanding the Surge in Large Language Model Use by Secondary Students
Project Overview
The document examines the increasing integration of large language models (LLMs), such as ChatGPT, in secondary education, noting that over 70% of students have adopted these tools for subjects like language arts, history, and math. While these technologies hold promise for enhancing learning, students report mixed experiences due to issues with accuracy and restrictions imposed by schools. The authors emphasize the necessity for LLMs designed specifically for educational purposes, aiming to improve their relevance and effectiveness in classroom settings. They advocate for the development of subject-specific models and the establishment of AI classrooms to facilitate better use of LLMs, ultimately seeking to enhance access and educational outcomes, particularly for underserved communities. This approach underscores the potential of generative AI to transform educational practices while addressing the challenges that accompany its implementation.
Key Applications
AI-assisted learning tools (e.g., LLMs, AI tutors)
Context: Secondary students from grades 7 to 12 across the United States, including underserved communities
Implementation: Surveys and proposals conducted to understand and implement AI-assisted tools like large language models (LLMs) and AI tutors trained on teaching materials to provide personalized learning experiences.
Outcomes: 70% of students reported using AI tools for academic assistance, leading to personalized learning experiences and potential inclusivity supporting diverse learning needs.
Challenges: Inaccurate responses from LLMs (hallucinations), mixed effectiveness across subjects, resources required for implementation, and potential acceptance issues from educators.
Implementation Barriers
Ethical/Policy Barrier
Schools have restrictive policies against LLM usage due to concerns over cheating, inaccuracies, and the need for responsible integration.
Proposed Solutions: Understand student usage patterns and develop curriculum to better incorporate LLMs responsibly.
Access Barrier
Students from resource-poor backgrounds may not have equal access to LLMs, exacerbating educational inequalities.
Proposed Solutions: Develop tailored LLMs for educational use and promote AI classrooms to provide equitable access.
Project Team
Tiffany Zhu
Researcher
Kexun Zhang
Researcher
William Yang Wang
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Tiffany Zhu, Kexun Zhang, William Yang Wang
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