Skip to main content Skip to navigation

Learnersourcing in the Age of AI: Student, Educator and Machine Partnerships for Content Creation

Project Overview

The document explores the role of generative AI in education, particularly through the concept of learnersourcing, where students collaboratively create educational content. It underscores the advantages of involving students in content creation, which fosters higher-order thinking and builds extensive repositories of educational materials. A framework is proposed for integrating AI into learnersourcing, encompassing content creation, evaluation, utilization, and instructor oversight. Key case studies, including PeerWise and RiPPLE, exemplify this framework, showcasing how generative AI can enhance both content generation and assessment quality. The findings indicate that while generative AI significantly improves engagement and learning outcomes, there are challenges related to potential biases and the quality of AI-generated content. Overall, the use of generative AI in education not only enriches the learning experience by empowering students to contribute but also necessitates careful consideration of its limitations and ethical implications in educational contexts.

Key Applications

Student-Generated Assessment Tools

Context: Web-based platforms allowing students to create, publish, and evaluate multiple-choice questions and other learning resources in various educational settings, including higher education in computer science courses.

Implementation: Students create multiple-choice questions and other educational resources that undergo a peer-review process. Instructors oversee the quality and provide analytics on student engagement and performance, ensuring a collaborative learning environment.

Outcomes: Over 7 million questions and 80,000 resources created by students, leading to increased engagement, investment in learning, and improved assessment performance.

Challenges: Quality control of student-generated content, ensuring adequate participation, maintaining quality evaluations, and managing instructor workload efficiently.

AI-Generated Programming Resources

Context: Used in computer science courses to enhance programming education by automatically generating programming exercises and explanations.

Implementation: Leveraging large language models to create and explain programming tasks, facilitating a more efficient learning experience for students.

Outcomes: Enhanced learning efficiency and understanding of programming concepts through AI-generated content.

Challenges: Dependence on the quality of AI-generated content and the need for human oversight to ensure accuracy and relevance.

Implementation Barriers

Implementation Barrier

The need for educators to balance the responsibilities of content creation with other academic commitments.

Proposed Solutions: Effective strategies for incentivizing students and training them in content creation need to be developed.

Quality Control Barrier

Challenges in maintaining the quality of student-generated content and potential overdependence on AI tools may lead to unoriginal or inaccurate content.

Proposed Solutions: Designing interfaces and scaffolds for AI tools to assist students effectively, implementing peer reviews, and instructor oversight to evaluate content quality.

Trust and Engagement Barrier

Students may not fully understand or appreciate the benefits of learnersourcing, leading to difficulty in motivating all students to participate in content generation.

Proposed Solutions: Explicitly discussing the advantages and rationale for including learnersourcing in the curriculum, incorporating gamification strategies, and incentivizing contributions.

Bias

Potential biases in peer evaluations and content generation.

Proposed Solutions: Providing training for students on unbiased assessment techniques and establishing clear guidelines.

Project Team

Hassan Khosravi

Researcher

Paul Denny

Researcher

Steven Moore

Researcher

John Stamper

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Hassan Khosravi, Paul Denny, Steven Moore, John Stamper

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

Let us know you agree to cookies