Skip to main content Skip to navigation

The AI Assessment Scale (AIAS) in action: A pilot implementation of GenAI supported assessment- A Preprint

Project Overview

The document examines the implementation of the AI Assessment Scale (AIAS) at British University Vietnam (BUV) to effectively integrate Generative AI (GenAI) into educational assessments. It emphasizes the advantages of GenAI in enriching learning experiences while simultaneously tackling issues related to academic integrity. The AIAS serves as a structured framework that categorizes the extent of AI integration in assessments, ranging from minimal to comprehensive use. Findings from a pilot study indicated a notable decrease in academic misconduct, an increase in student performance, and a positive transformation in teaching methodologies. The document underscores the importance of utilizing GenAI responsibly within educational contexts, advocating for a balanced approach that maximizes its benefits while minimizing potential challenges.

Key Applications

AI Assessment Scale (AIAS)

Context: Higher education, specifically at British University Vietnam, targeting educators and students.

Implementation: The AIAS was developed to provide a flexible framework for integrating GenAI into assessments, allowing varying levels of AI assistance.

Outcomes: A 5.9% increase in student attainment, a 33.3% increase in module passing rates, and a significant reduction in academic misconduct cases.

Challenges: Initial resistance to GenAI tools, the need for policy adjustments, and the challenge of ensuring fair use across different student demographics.

Implementation Barriers

Policy

Initial bans on GenAI tools due to concerns about academic integrity.

Proposed Solutions: The implementation of the AIAS framework to guide ethical use and integration of GenAI in assessments.

Technological

Limitations of AI detection tools in identifying AI-generated content.

Proposed Solutions: The AIAS encourages a focus on ethical engagement with GenAI rather than solely relying on detection methods.

Project Team

Leon Furze

Researcher

Mike Perkins

Researcher

Jasper Roe

Researcher

Jason MacVaugh

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Leon Furze, Mike Perkins, Jasper Roe, Jason MacVaugh

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