Artificial intelligence has come a long way in recent years, and post-OpenAI development is no exception. GPT-4 chatbot is a powerful tool that uses machine learning to generate natural language text. He has achieved remarkable feats such as passing the bar exam and several other standardized tests with flying colors. However, new research shows that AI still cannot match human intelligence in accounting.
Researchers at Brigham Young University and 186 other universities tested OpenAI’s ChatGPT chatbot prototype. The exam includes more than 25,000 accounting practice tests of various types and difficulties, including areas such as financial accounting, auditing, taxation, and more. The results were eye-opening: While ChatGPT’s performance was impressive, students outperformed the AI in the overall score.
Although the ChatGPT score is lower than the student, it can perform many types of questions. He scored higher than the average student in several assessments, especially in accounting and auditing information systems. However, he struggled with other areas such as taxation and management accounting. One possible explanation for this is that ChatGPT is having difficulty with the mathematical processes required for these queries.
Despite ChatGPT’s impressive capabilities, the study’s authors note that it is not a perfect tool. Students outperformed AI, especially in higher-order questions that require more critical thinking and problem-solving skills. He also struggles with short answer questions and sometimes gives wrong answers or multiple answers to the same question.
In conclusion, while ChatGPT is an interesting development, it is still not compatible with the human mind when it comes to accounting. It can be a useful tool for some tasks, but students should not rely on it to replace traditional learning methods. Like all AI tools, they have limitations and must be used in conjunction with human intelligence and critical thinking skills.