Study Finds ChatGPT Bombs Accounting Class | Accounting Today
A recent study from Brigham Young University found that ChatGPT, an AI chatbot, failed miserably in an accounting class, with a score of only 47.4%, which is not even a passing grade. The study involved 327 co-authors from 186 educational institutions in 14 countries who contributed 25,181 classroom accounting exam questions. Additionally, undergrad BYU students provided another 2,268 textbook test bank questions to ChatGPT. The questions covered auditing, financial accounting, managerial accounting, and tax, and varied in difficulty and type; true/false, multiple choice, short answer, etc.
Human students did much better, averaging 76.7%. The AI performed better than humans in only 11.3% of questions, mainly on accounting information systems (AIS) and auditing. However, it scored worse than humans on tax, financial, and managerial assessments, possibly because it is made for language rather than math.
During testing, ChatGPT sometimes failed to recognize mathematical operations and made nonsensical errors like adding two numbers in a subtraction problem or dividing numbers incorrectly. Nevertheless, opportunities to cheat have always existed. That said, researchers focused on what they could do with this technology to improve the teaching process for faculty and the learning process for students.
ChatGPT failed to pass an accounting class, which raises the question of whether or not it can pass the CPA exam. Accounting Today is currently exploring this and plans to release their findings soon.
In a practice CPA exam, ChatGPT scored poorly, receiving 39% in REG, 46% in AUD, 35% in FAR, and 48% in BEC. The Senate Finance Committee recently examined the Tax Cuts and Jobs Act's impact on drug makers' taxes and found that the law allowed some of the biggest companies to lower their tax bills by shifting profits abroad.
Accounting Today provides breaking news, in-depth features, and various resources and services to the accounting community.
