This research aims to examine legal liability in cases of copyright infringement by Artificial Intelligence (AI) in Indonesia and to evaluate the clarity of the existing legal mechanisms. Based on the analysis, it is found that Law Number 28 of 2014 on Copyright (UUHC) still adheres to an anthropocentric legal paradigm that recognizes only human beings or legal entities as legitimate legal subjects and creators. In the event of copyright infringement by AI, legal responsibility is transferred to the users, developers, or owners of the system, even though there is no explicit and comprehensive normative instrument available to determine liability based on intent, negligence, or economic gain. Furthermore, although institutional structures such as the Directorate General of Intellectual Property (DJKI), the Commercial Court, and procedures for litigation and ADR are already in place, the copyright legal system in Indonesia is not yet fully adaptive to the complexity of AI-based infringements—particularly regarding algorithmic processes and the use of unauthorized training data. The potential for copyright violations by AI encompasses various forms, including unauthorized reproduction, illegal compilation, covert plagiarism, infringement of distribution rights, and moral rights violations. The lack of transparency in training data and algorithmic traces exacerbates attribution challenges and creates systemic inequality between technological corporations and the vulnerability of individual creators. Therefore, it is recommended that the government design legal reforms to explicitly recognize AI-generated works, apply the principle of accountable innovation, mandate algorithmic transparency, strengthen the roles of DJKI and Collective Management Organizations, and establish compensation funds for creators. This study highlights the urgency of reconstructing copyright law’s substantive aspects to ensure justice and effective protection in the age of artificial intelligence.