Legal Liability for Artificial Intelligence-Generated Content from an Intellectual Property Perspective in Indonesia
TL;DR
The findings show that artificial intelligence can be neither a legal subject nor an author and formulates a tiered liability model distributing the burden among model developers, platform providers, and users according to degree of control, creative contribution, knowledge of risk, and discharge of the duty of care.
Abstract
Generative artificial intelligence now produces economically valuable works on a massive scale, while Indonesian copyright law still recognises human beings as the only possible authors. This study analyses the legal status of artificial intelligence-generated content from an intellectual property perspective and formulates a liability model suited to the Indonesian legal system. This normative legal research applies statutory, conceptual, case, and comparative approaches. Primary legal materials comprise Law Number 28 of 2014 on Copyright, the Indonesian Civil Code, and the Electronic Information and Transactions Law, while secondary materials consist of reputable literature published within the last five years, analysed through content analysis, legal interpretation, and deductive syllogism. The findings show that artificial intelligence can be neither a legal subject nor an author. Copyright protection attaches only insofar as a human creative contribution determines the final form of the work, consistent with the direction and supervision logic of Article 34. Liability for infringement must therefore be transferred to the human or legal entity connected to the system. This study formulates a tiered liability model distributing the burden among model developers, platform providers, and users according to degree of control, creative contribution, knowledge of risk, and discharge of the duty of care.