<![CDATA[The rapid development of autonomous artificial intelligence (AI) systems through machine learning and self-learning algorithms has challenged the traditional fault-based civil liability framework in Indonesia. Classical liability constructs assume that harm can be traced to human intent or negligence, yet autonomous AI can generate unpredictable outcomes without direct human control. This normative juridical study aims to analyze the inadequacy of Indonesian legal norms in attributing responsibility for damages caused by autonomous AI systems and to formulate a more adaptive liability reconstruction model. The research applied a statutory and conceptual approach, examining Articles 1365–1367 of the Civil Code, Law No. 8 of 1999 on Consumer Protection, Law No. 11 of 2008 on Electronic Information and Transactions, and relevant scholarly literature. Data were analyzed through qualitative content analysis, comparing Indonesian legal provisions with international risk-based liability models, and reconstructing a coherent framework for high-risk autonomous AI. The findings reveal that fault-based liability is insufficient for AI ecosystems involving multiple actors and probabilistic decision-making, creating gaps in accountability and victim protection. The study proposes a risk-based liability model integrating strict liability for high-risk AI, joint and several liability, mandatory insurance, burden of proof reversal, and algorithmic transparency. This model is expected to enhance legal certainty, strengthen victim protection, and provide a regulatory foundation for responsible AI governance in Indonesia.]]>
Copyrights © 2026
