Epistemologies of Authentic Assessments Amidst the AI Advancement: An Evaluation of the Divergence Between Traditional and AI-Generated Scientific Illustrations

Abstract

As artificial intelligence (AI) continues to advance and improve its capacity to provide detailed explanations, a critical question emerges: Can students still cultivate a deep and meaningful understanding of educational material? This question underscores the need to reassess and innovate the methods that researchers use to evaluate student comprehension in this rapidly evolving technological landscape. This study focuses on exploring the significant differences between traditional scientific illustrations, crafted by students and those generated by AI tools. A diverse group of 267 students from the University of Venda in South Africa participated in the evaluation process, which used a comprehensive assessment tool designed to allocate a maximum of 10 marks based on various criteria, such as clarity, accuracy, and relevance of the illustrations. To assess the capabilities of different AI tools, the researchers specifically instructed the selected students to use the AI system to generate an illustration of the root cap structure of the buttercup plant. Through this approach, students were empowered to effectively monitor and reflect on their own learning processes and outcomes. The findings from this study suggest that educational instructors should emphasise the importance of critical evaluation when students engage with AI-generated content. In particular, educators and lecturers should guide students to assess AI responses by scrutinising factors such as accuracy, potential biases inherent in AI algorithms, and the degree of simplification presented in the illustrations. By promoting an active engagement with technology, students are encouraged to become discerning users of AI tools, rather than passive consumers of content. State the contribution of this study to scholarship. Keywords: Active learning, Artificial Intelligence, Balance, practical application, digital competence, critical thinking, self-supervised learning.

Ähnliche Arbeiten