This paper explores the impact of generative artificial intelligence on classification models by implementing a comprehensive pipeline in Python. The primary goal is to analyze the performance of classification algorithms on an original dataset and an augmented dataset generated using two different synthetic data generation methods. This research includes data preprocessing, application of four machine learning classification methods, evaluation of classification performance, and synthetic data generation using generative AI techniques. The results of classification performance on both the original and generated datasets are evaluated using metrics such as accuracy, precision, recall, and F1-score, and compared. The findings show that synthetic data generation methods, particularly SMOTE, led to significant improvements in model performance across all datasets, enhancing recall and overall classification accuracy. Furthermore, classifiers trained on the combined datasets, which included both original and synthetic data, performed comparably or better than those trained solely on the original data, demonstrating the effectiveness of data augmentation. These results underline the potential of using generative AI techniques to address challenges posed by imbalanced datasets and improve the robustness of machine learning models in various real-world applications.