ChatGPT underperforms students in accounting exams and struggles with math process

Researchers found that students performed better on accounting exams than ChatGPT, OpenAI’s chatbot.

Despite this, they said ChatGPT’s performance was “impressive” and that it was a “game changer that will change the way everyone teaches and learns – for the better”. Researchers from Brigham Young University (BYU), USA, and 186 other universities wanted to know how OpenAI’s technology would perform in accounting exams. They published their findings in the journal Issues in Accounting Education.

On the Researchers Accounting exam, students scored an overall average of 76.7%, compared to 47.4% for ChatGPT.

While in 11.3% of questions, ChatGPT scored above the student average, particularly well on Accounting Information Systems (AIS) and Auditing, the AI ​​bot performed worse on fiscal, financial and managerial assessments. The researchers believe this could be because ChatGPT struggled with the mathematical processes required for the latter type.

The AI ​​bot, which uses machine learning to generate natural language text, was also found to perform better on true/false questions (68.7% correct answers) and multiple-choice questions (59. 5%), but struggled with short-answer questions (between 28.7% and 39.1%).

In general, the researchers said it was more difficult for ChatGPT to answer higher-order questions. In fact, ChatGPT has sometimes been found to provide authoritative written descriptions for incorrect answers or answer the same question in different ways.

They also found that ChatGPT often provided explanations for its answers, even if they were incorrect. Other times, he would select the wrong multiple-choice answer, despite accurate descriptions.

The researchers mostly noted that ChatGPT sometimes made things up. For example, when providing a reference, it generated a real-looking reference that was completely fabricated. The work and sometimes the authors did not even exist.

The bot has also been seen to make silly math mistakes such as adding two numbers in a subtraction problem or incorrectly dividing numbers.

Wanting to add to the intense ongoing debate about how models like ChatGPT should be considered in education, the study’s lead author David Wood, an accounting professor at BYU, decided to recruit as many professors as possible to see how the AI ​​fared against real college accounting students. .

Its co-author recruitment pitch on social media exploded: 327 co-authors from 186 educational institutions in 14 countries participated in the research, contributing 25,181 classroom accounting exam questions.

They also recruited BYU undergraduates to feed another 2,268 textbook test bank questions to ChatGPT. The questions covered AIS, auditing, financial accounting, management accounting and taxation, and varied in difficulty and type (true/false, multiple choice, short answer).

---