Abstract
In the context of the accounting professional's problems and tasks that can be optimized with Artificial Intelligence applications, the research aimed to identify and evaluate the performance of the ChatGPT tool, model GPT-4.0, in resolving questions from the Accounting Sufficiency Exam, as a proxy for problems in accounting activity. For this verification, we applied questions from four editions of the Exam in ChatGPT 4.0. The results indicate approval in the four editions in which only 20 to 23% of candidates were approved, while the detailed performance showed 71% of ChatGPT's correct answers compared to 44% of the candidates.
Keywords: Artificial Intelligence, ChatGPT, Tasks.