ChatGPT Forces Universities To Adapt Or Retreat
How did AI-generated submissions compare to human students in detection?
AI-generated submissions were found to be more difficult to detect compared to human students. In a study conducted by the University of Reading, 94% of AI-generated submissions went undetected by markers. Additionally, AI consistently achieved higher grades than human students, except in third-year exams requiring more abstract reasoning.
What assessment alternatives does Tim Mousel suggest?
Tim Mousel suggests several assessment alternatives to traditional essays, including project-based assessments, problem-solving scenarios, oral presentations or debates, collaborative group projects, reflective journals or portfolios, experiential learning assignments, peer teaching or tutoring, creation of original content, interactive simulations or role-playing, and open-ended research projects. These methods aim to foster critical thinking, creativity, collaboration, and communication skills, which are less easily replicated by AI.
What are Dr. Jennifer Chang Wathall's views on assessment practices?
Dr. Jennifer Chang Wathall advocates for a complete overhaul of assessment practices, emphasizing the need to focus on the process of learning and growth over time. She believes assessment should celebrate individual strengths and talents, and be a continuous process measured against qualitative criteria. This shift would move the focus from single high-stakes exams to a more holistic evaluation of a student's learning journey.