The use of Artificial Intelligence (AI) in the public sector offers significant improvements in bureaucratic effectiveness but simultaneously raises serious concerns regarding data privacy, security, and algorithmic bias. In response, the Ministry of Communication and Digital (MoCD) issued Regulation No. 5 of 2025 concerning Public Electronic System Operators to address these challenges. This research aims to analyze the juridical status of data classification under this regulation as a foundational element for establishing safe and ethical AI governance in the Indonesian public sector. This research employs a normative legal research method using statutory and conceptual approaches, with qualitative and prescriptive analysis of secondary legal materials, including primary and secondary legal sources. The findings reveal that MoCD Regulation No. 5 of 2025 provides legal safeguards through data classification into three categories: open data, restricted data, and confidential or closed data. These categories serve as essential parameters for AI algorithms when processing public information. The juridical analysis confirms that data classification is not merely an administrative measure but constitutes a conditio sine qua non—an absolute necessity—for mitigating the risks of data leaks and algorithmic bias in government-level AI systems. The study concludes that harmonizing this regulation with Law No. 27 of 2022 concerning Personal Data Protection, commonly referred to as the Personal Data Protection Law, is essential to creating a comprehensive legal framework that ensures AI implementation in the public sector adheres to the principles of data security, privacy protection, and digital sovereignty. This research recommends the development of detailed technical guidelines to operationalize data classification in AI systems and the alignment of consent mechanisms with the Personal Data Protection Law to address the asymmetrical power relationship between the state and citizens regarding the use of public data for AI training.
Copyrights © 2026