New Framework Aims to Guide AI Use in Assessments
Educator Leon Furze has introduced the AIAS (AI Assessment in Schools) framework to guide responsible AI use in student evaluations. The framework moves beyond simple allow/disallow policies by breaking down assessments into components like planning, research, and editing. It is designed to facilitate conversations between educators and students about appropriate AI use in different contexts.
- The AIAS framework was co-developed by Mike Perkins, Leon Furze, Jasper Roe, and Jason MacVaugh and was first introduced in 2023 before being updated in 2024. It is designed to provide clear guidance on the acceptable use of generative AI in assessments. - The framework consists of five distinct levels of AI integration: No AI, AI Planning, AI Collaboration, Full AI, and AI Exploration. These levels range from completing tasks entirely without AI in a controlled environment to using AI creatively to solve problems and co-design new assessment approaches. - A key motivation for creating the AIAS was to move beyond simple "cheating" narratives surrounding AI and to provide a more nuanced way for educators to articulate their decisions about AI use to students. It addresses the ineffectiveness of AI detection tools by focusing on transparent assessment design. - The AIAS is grounded in social constructivist principles and is intended to facilitate dialogue between educators and students about appropriate AI use. Its flexible design allows for adaptation across different educational contexts, including K-12 and higher education. - A pilot study of the AIAS at British University Vietnam indicated a significant reduction in academic misconduct cases related to generative AI. - The framework has been recognized by the Australian Tertiary Education Quality and Standards Agency (TEQSA) as a tool to assist with implementing generative AI in assessment. - AI-driven assessments offer the potential for personalized learning by tailoring content to individual student needs and providing real-time feedback. This can help to improve student outcomes and engagement. - Significant ethical challenges in AI-driven assessment include the potential for algorithmic bias, data privacy and security concerns, and a lack of transparency in how AI systems arrive at their conclusions.
