Implementation of a Generative AI Assistant in K-12 Education: The CyberScholar Initiative
Project Overview
The document examines the use of CyberScholar, a generative AI assistant developed to enhance writing skills in K-12 education by providing personalized feedback across subjects like English Language Arts, Social Studies, and Modern World History. By utilizing prompt engineering and leveraging a large language model, CyberScholar aims to deliver tailored responses that support students in their writing endeavors. The findings indicate that the tool has the potential to improve both student writing proficiency and teacher efficiency, making it a valuable resource in the educational landscape. However, the implementation of CyberScholar is not without challenges, as it raises ethical concerns, presents user interface difficulties, and exhibits risks of algorithmic bias. Despite these hurdles, the overall outlook for generative AI in education is optimistic, with CyberScholar serving as a noteworthy example of how technology can be harnessed to support learning outcomes.
Key Applications
CyberScholar - Generative AI writing assistant
Context: K-12 education in various subjects including English Language Arts, Social Studies, and Modern World History for students in grades 7, 8, 10, and 11.
Implementation: The tool was piloted in four schools, utilizing prompt engineering and fine-tuning of a Large Language Model based on teachers' rubrics and a customized corpus.
Outcomes: Enhanced student writing abilities, provided timely feedback, and reduced teachers' workload by automating feedback processes.
Challenges: Concerns about ethics, algorithmic bias, and technical issues with the user interface were noted.
Implementation Barriers
Ethical and Algorithmic Bias
Concerns about verifying AI-generated work, potential disruption to traditional teaching practices, and inherent biases in AI that may affect feedback quality and student learning outcomes.
Proposed Solutions: Creating ethical conditions for AI use in classrooms, developing AI literacy among students, and implementing filters to identify and flag biases in AI-generated feedback.
Technical
User interface issues and technical glitches that hinder usability, especially on school devices.
Proposed Solutions: Improvements to the user interface based on feedback and making the tool more intuitive.
Project Team
Vania Castro
Researcher
Ana Karina de Oliveira Nascimento
Researcher
Raigul Zheldibayeva
Researcher
Duane Searsmith
Researcher
Akash Saini
Researcher
Bill Cope
Researcher
Mary Kalantzis
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Vania Castro, Ana Karina de Oliveira Nascimento, Raigul Zheldibayeva, Duane Searsmith, Akash Saini, Bill Cope, Mary Kalantzis
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