One profession that AI cannot outperform humans

Weather of Arabia - so far, artificial intelligence has shown ingenuity in telling stories and writing texts creatively, and often providing human-like answers, even outperforming some doctors in caring for and empathizing with patients, which raised concerns about its unlimited ability that could be a cause In the dispensation of some sectors with humans, however, there is one profession that artificial intelligence has failed to find a foothold in, which makes the people who are appreciative of it less nervous.

The " Fox Business " network quoted the results of a study that concluded that artificial intelligence is weak in mathematics and mathematical operations, which keeps competition away from workers in the accounting task.

This comes as the AI model ChatGPT - the fastest growing and most popular AI platform to date - has raised concerns about its ability to help students cheat on coursework and test materials. The robot passed 13 out of 15 British Higher School Examinations, or AP exams, and achieved a near-perfect score on the Graduate Record Examination (GRE).

But the new study, led by Brigham Young University accounting professor Dr. David Wood, finds that AI platforms often struggle to understand mathematical operations, and often embellish data to cover up errors when they happen.

Wood's study sought to test ChatGPT's proficiency in completing accounting exams compared to actual accounting students. 25,181 questions were submitted on information systems, auditing, financial accounting, management accounting, and taxation from 186 educational institutions in 14 countries.

The questions were presented in various formats with varying levels of difficulty, with a combination of multiple choice, true/false, and written response prompts being used.

The study found that students scored higher than ChatGPT, outperforming the chatbot by more than 30% - an average of 76.7% versus 47.4% for ChatGPT.

ChatGPT only outperformed students in 11.3% of the questions, particularly those related to auditing and accounting information systems. The chatbot was also more adept at answering multiple-choice and true/false questions, scoring 59.5% and 68.7% in each format separately - however, it performed much lower on short-answer questions, where it got at between 28.7% and 39.1%.