This paper explores the transformative impact of generative artificial intelligence (AI) on the audiovisual sector, with a particular focus on text-to-image and text-to-video technologies. Using models like Sora, it shows how multimodal AI can now produce increasingly coherent synthetic content. These developments are reshaping creative workflows by redefining the roles of human creators and expanding the possibilities for artistic expression, automation, and experimentation. Drawing on recent scholarly literature and practical experimentation, the study highlights both the technical achievements and persistent limitations of generative models. While users can now produce visually compelling results through prompt-based interaction, challenges remain (especially in video generation) with inconsistencies in spatial continuity, object coherence, and semantic alignment. The paper critically assesses prompt engineering as a creative skill, underscoring its relevance in achieving professional-grade outcomes. The discussion also addresses pressing ethical concerns, particularly the rise of deepfakes and the resulting erosion of public trust in audiovisual media. The rise of accessible generative tools raises urgent concerns about authorship and authenticity. This study advocates for a balanced approach, one that fosters innovation while ensuring ethical accountability, technical literacy, and the development of appropriate regulatory frameworks. Ultimately, the research emphasizes the need to reframe the landscape of audiovisual production through a critical, interdisciplinary lens, recognizing AI’s potential to create and disrupt, democratize, and redefine creative authorship.