Abstract: This study explores the influence of GAN-generated deepfakes within the creative media industry by applying a hybrid framework that integrates the Naïve Bayes algorithm and the Expectation Confirmation Model (ECM). As deepfake technologies become increasingly prevalent in filmmaking, digital art, and online content, they raise both creative opportunities and ethical concerns. The research employed a qualitative exploratory design involving 25 purposively selected participants from the Philippine creative sector, all of whom had experience with digital content production. Using a Naïve Bayes classifier, participant responses were analyzed for sentiment trends, revealing a majority of positive or neutral perceptions when deepfakes were applied ethically and creatively. The ECM component provided behavioral insights, demonstrating that user satisfaction was highest when the perceived performance of deepfake content met or exceeded initial expectations. Conversely, negative disconfirmation, often linked to ethical discomfort or perceived misuse, led to dissatisfaction and rejection of the technology. The findings emphasize the importance of managing audience expectations and embedding ethical standards into creative workflows involving synthetic media. By combining computational sentiment analysis with user-centric evaluation, this study contributes a scalable, interdisciplinary framework for assessing the reception and impact of AI-driven media innovations.