Abstract: This study explores the impact of deepfake media on the creative and entertainment industry, with insights drawn from GAN-based deep learning models. Deepfakes offer powerful creative potential, enabling digital resurrection of performers, virtual concerts, and enhanced storytelling techniques. At the same time, their hyperrealism raises ethical concerns around consent, authorship, and audience transparency. Using the Diffusion of Innovations theory, the research finds that deepfakes are transitioning from early adoption to broader use among content creators and media professionals, while public acceptance remains cautious due to issues of trust and misuse. The same qualities that make deepfakes compelling in entertainment also make them dangerous tools for misinformation, fraud, and non-consensual content. As such, the study highlights the need for industry guidelines, public education, and updated legal frameworks to balance innovation with protection. Ultimately, the future of deepfake media in entertainment hinges on responsible integration and proactive safeguards to ensure its creative promise does not outweigh its societal risks.