In this study, Byzalov and approximately 300 fellow researchers investigated ChatGPT’s performance on various exam types and conditions, comparing it to student performance. The study found that while students generally outperformed ChatGPT, the AI showed comparable results in multiple-choice and True/False questions.This research highlights the current strengths and limitations of AI in academic assessments. It ultimately suggests that ChatGPT (version 3), performs alright on certain question formats, but human students still demonstrate a higher overall level of understanding and performance.
Balancing the Books: How Does ChatGPT Handle Accounting?
