A Toolkit for Sustainable Educational Environment in the Modern AI Era: Guidelines for Mitigating the Misuse of GenAI in Assessments

Abstract

Large Language Models (LLMs), utilised through a number of software tools, are now widely used by students to support their learning and complete their assignments. Evaluating the intellectual abilities of learners has always been an important component of assessment to gauge learning performance and learning outcomes, an aspect that has become challenging with the availability of LLMs. Nevertheless, the requirements of digital evaluation remain an important mechanism to support equality and integrity in academia, especially for authentic assessments. Educators require a methodology to evaluate the reliability of evaluating student performance; hence, Artificial Intelligence (AI) can also be used to guide assessment design methods. This study introduces a toolkit for a sustainable educational environment that leverages AI to mitigate the misuse of AI by students in completing their assignments and assessing their learning, while empowering professional degrees, such as engineering, to maintain their accreditation status. This article features views of prominent academics, researchers, distinguished scientists and professional services staff from various disciplines within Queen Mary University of London (QMUL), along with external national and international experts from both academia and industry, who discuss effective guidelines and learning practices to support fair assessment and reduce the misuse of AI by students in completing their assessments.