Should We Create Better Assignments Instead of Worrying about Generative AI (Such as ChatGPT)?

Abstract

This study critically examines the impact of generative AI on university assessments, questioning whether current assignments develop the practical skills necessary for the workforce. Analyzing final exams from five universities, the research identifies an overreliance on multiple-choice questions, comprising 70% of assessments, and highlights the potential for AI-assisted cheating in writing tasks. Findings reveal a significant misalignment between the skills evaluated in academic assignments and those demanded in professional settings, underscoring the need for innovative assessment methods that foster authentic learning and skill application. The study advocates for the integration of AI tools into educational practices to enhance academic integrity and student preparedness for real-world challenges.