The Future of Copyright Protection for AI-Generated Art: Lessons from the Ghiblification Phenomenon

The rapid development of AI-generated art that mimics the signature styles of human artists, exemplified by the ghiblification phenomenon, poses a crucial problem for the copyright protection of artistic style, which is traditionally unprotected. This research aims to critically analyze this legal challenge, conduct a comparative study of the copyright regulatory frameworks in Indonesia, the United States, the European Union, and Japan concerning AI art and the issue of artistic style, and formulate policy recommendations for Indonesia. Using a normative juridical research method through statute, conceptual, and comparative approaches, this study examines the legislation, doctrine, and practices in these four jurisdictions. The results show that ghiblification confirms the vulnerability of artistic style; although style as an idea is not protected, the replication of specific expressions by AI can still potentially constitute copyright infringement. The comparative analysis reveals significant variations in approach: the United States strictly requires human authorship, Japan offers flexibility for using data for AI training with a proviso, and the European Union seeks a balance through a TDM exception, while Indonesia still faces a specific regulatory vacuum. Nevertheless, a global consensus exists on the importance of human creative contribution for the recognition of copyright. It is concluded that the existing copyright legal framework, particularly in Indonesia, is inadequate to respond to the disruption of AI-generated art, thus requiring urgent juridical adaptation. This study recommends legal reform in Indonesia, including the clarification of the status of AI-generated art, the consideration of licensing models, and the strengthening of moral rights aspects in order to balance technological innovation with fair artistic protection.