Copyright in the Era of Generative AI: Mapping Legal Gaps in Indonesian Copyright Protection

Abstract

The rapid advancement of artificial intelligence (AI), particularly generative and data-driven models, has introduced unprecedented challenges to traditional copyright frameworks. This study examines how the technical workflow of AI from data input, text and data mining (TDM), and feature extraction to model building, assessment, and output generation intersects with legally protected creative expressions. By mapping these technical stages against Indonesia’s Copyright Law (Law No. 28/2014), the research identifies five critical areas of concern: fair use in TDM, non-literal copying, AI authorship, liability for AI-generated infringements, and dataset transparency. Using a doctrinal legal research approach, the analysis reveals significant normative gaps within the current copyright regime, which remains anthropocentric and ill-equipped to regulate autonomous or algorithmic forms of creativity. The legal gap map developed in this study demonstrates that Indonesian copyright law lacks explicit provisions governing machine-based data processing, algorithmic replication of styles, ownership of AI-generated outputs, allocation of responsibility among AI ecosystem actors, and mandatory disclosure of training datasets. These regulatory deficiencies create substantial uncertainty for creators, AI developers, platforms, and users, ultimately weakening the enforcement of copyright in digital contexts. The study concludes that comprehensive legal reform—potentially through updated exceptions, sui generis protection, and clearer governance of AI datasets—is essential for ensuring that Indonesia’s copyright system remains effective, balanced, and future-ready in the age of artificial intelligence.