Regulasi Kecerdasan Buatan Tanpa Sanksi: Analisis Kritis Rancangan Peraturan Presiden (RPerpres) tentang Etika Kecerdasan Artifisial dalam Perspektif Hukum Indonesia

Abstract

The rapid development of artificial intelligence (AI) in Indonesia has been accompanied by a rising number of AI-based digital fraud cases, including deepfake and voice cloning, while Indonesia still lacks a dedicated law for AI. In response, the government drafted a Presidential Regulation on AI Ethics (RPerpres Etika AI), adopting a soft law approach without sanctions, which differs from the European Union's risk-based approach in the EU AI Act that includes binding sanctions. This study is a prescriptive-analytical normative legal research using statutory, comparative, and conceptual approaches, aiming to analyze the normative construction of the RPerpres and to assess the adequacy of its soft law approach without sanctions in addressing the risk of AI misuse. The findings show that the RPerpres establishes nine ethical values, distributes responsibility among Users, Sector Actors, and Ministries/Agencies, and classifies AI risk into three levels, but contains no binding sanctions, lacks clarity on enforcement authority, and remains dependent on the ITE Law, the PDP Law, and the Criminal Code, none of which were specifically designed for the characteristics of AI. Compared to the EU AI Act, which applies risk classification, mandatory conformity assessment, and tiered administrative sanctions, the RPerpres risks becoming a regulation without substantive binding force. This study identifies several aspects that need to be strengthened, such as strengthening through risk-based sanctions, and consolidation through a dedicated AI law, in order to provide stronger legal certainty while remaining proportionate to Indonesia's institutional and economic context.