Core Concepts
Effective assessment design and evaluation strategies are crucial in the era of Generative AI to ensure academic integrity, authentic learning, and responsible use of AI tools.
Abstract
The paper discusses the key challenges and opportunities presented by Generative AI (GenAI) in the education ecosystem, particularly in the context of assessment design and evaluation. It proposes a classification of assessments into two main types: Learning-based Assessments and Action-based Assessments.
Learning-based Assessments focus on measuring and recalling factual concepts, with or without the use of GenAI tools. These assessments are designed to ensure learning of subject matter, problem formulation skills, and understanding of GenAI tools. Action-based Assessments, on the other hand, emphasize the application of knowledge and skills using GenAI tools to solve real-world problems.
The paper outlines a comprehensive assessment design process that involves determining the assessment types, the use of GenAI, applying authentic assessment design techniques, validating assessments against GenAI tools, and clearly communicating expectations to all stakeholders. Additionally, it discusses assessment evaluation strategies, including the traditional approach for Learning-based Assessments and the potential for automation using GenAI for Action-based Assessments.
The key strategies highlighted in the paper aim to ensure academic integrity, authentic learning, and the responsible use of GenAI tools in the educational context. The authors emphasize the importance of fostering critical thinking, creativity, problem-solving, and other essential graduate attributes, while also equipping students with the knowledge and skills to effectively utilize emerging GenAI applications.
Stats
"The two principles to be ensured during assessment design in the era of GenAI are: 1) Students should be able to use GenAI tools ethically and actively by recognising the ethics, limitations, biases, and implications of AI, and 2) Ensure student learning by using a variety of inclusive and contextualised approaches on assessments."
Quotes
"The primary contributions of this paper include assessment classification, assessment design process and evaluation strategies ensuring that the students' learning outcomes are preserved in the era of GenAI."
"The key in this step is to break down the task into actionable steps and focus on observable behaviours. Then, the criteria need to indicate what excellent, good, or bad performance on these actionable steps looks like."
"The assessment can be designed in staged assessments covering either one semester or multiple semesters. More weighting should be given to the process involved than the actual final product."