Evaluating AI-Assisted Boolean Expression Simplification: A Small-Scale Study

Abstract Artificial intelligence (AI) is increasingly being used to assist with mathematical, logical, and engineering-related tasks. In digital logic, Boolean expression simplification is important because it reduces unnecessary logical operations while preserving the original logical function. This study evaluates the performance of ChatGPT in simplifying five selected Boolean expressions with different levels of complexity. The test set consisted of two basic expressions, two intermediate expressions, and one complex expression. A standardized prompt was used for all five expressions. The generated solutions were evaluated based on final-answer correctness, Boolean-law identification, transformation validity, and logical equivalence. The results showed that ChatGPT correctly simplified all five selected expressions, resulting in an observed final-answer accuracy of 100%. All five transformations were also found to be valid, resulting in an observed transformation validity of 100%. However, one response contained inaccurate or inconsistent Boolean-law labeling, resulting in an observed Boolean-law identification accuracy of 80%. Truth-table verification confirmed that all five AI-generated simplified expressions were logically equivalent to their corresponding original expressions. These findings indicate that ChatGPT successfully performed the selected Boolean simplification tasks under the conditions of the study, while also demonstrating that AI-generated explanations should be independently checked, particularly for the accuracy of individual Boolean-law labels. Because the study used only five selected expressions and one AI system, the results should not be interpreted as evidence of general or universal accuracy in Boolean-expression simplification.