Indonesia - Ethical Guidelines for AI

Overview

The Konsep Pedoman Etika Kecerdasan Artifisial is a concept draft published by the Ministry of Communication and Digital (Komdigi) in August 2025 for public consultation. It accompanies a national AI White Paper (‘Buku Putih Peta Jalan Kecerdasan Artifisial Nasional’) and is presented as a values-driven, risk-based framework intended to guide the development, deployment, and governance of AI across public and private sectors in Indonesia. The draft builds on the foundation set by Ministerial Circular/SE No.9 of 2023 on AI ethics and signals the Ministry’s intention to move from guidance toward a binding Presidential regulation. The public consultation opened on 8 August 2025 and was extended to 29 August 2025 to gather broader stakeholder input. Extension notice.

Definitions

The draft defines key concepts in alignment with existing Indonesian instruments: ‘Artificial Intelligence / Kecerdasan Artifisial’ (programming capable of data processing, ML, NLP, neural networks), ‘Penyelenggara Sistem Elektronik (PSE)’, ‘Pelaku Usaha’ (AI programming actors), ‘High-Risk AI’ (applications with elevated potential for harm), and related technical terms such as ‘model provenance’, ‘human-in-the-loop’, and ‘adversarial robustness’. These definitions are intended to interoperate with the statutory definitions in Law No. 27/2022 on Personal Data Protection and existing electronic systems law and guidance. The draft aims to align legal and ethical terminology to reduce ambiguity across implementing agencies.

Governance and Institutional Framework

The draft places primary stewardship with the Ministry of Communication and Digital but envisages a multi-agency governance architecture involving sectoral regulators (e.g., OJK for finance, Ministry of Health for healthcare), national cybersecurity authorities (e.g., BSSN), and consumer protection bodies. It proposes a national coordination mechanism to: (1) classify high-risk systems; (2) maintain registries or notification schemes for significant AI deployment; (3) coordinate conformity assessment and market surveillance; and (4) harmonize sectoral rules to avoid regulatory fragmentation. The model is consultative and iterative — Komdigi’s consultation document and the White Paper were produced by a broad taskforce of government, academics, industry and civil society representatives to ensure cross-sectoral buy-in. Komdigi announcement.

Key Focus Areas

The draft’s focus areas include: (a) Governance & oversight (clear institutional roles, registries and inter-agency coordination); (b) Risk management (mandatory risk assessments and classification of high-risk use-cases); (c) Safety testing & evaluation (pre-deployment testing, adversarial resilience and continuous monitoring); (d) Transparency & disclosure (model cards, user notices, provenance data and AI-generated content labelling); (e) Data protection & privacy (privacy-by-design, DPIA alignment with Law No.27/2022); (f) Fundamental rights (anti-discrimination, due process protections for automated decisions); (g) Cybersecurity & model security (requirements for integrity, access controls and incident reporting); (h) Conformity assessment & registration (third-party audit pathways and potential mandatory registration for certain categories); (i) Accountability & documentation (retention of logs, audit trails and explainability records); and (j) Liability, redress & enforcement (administrative sanctions and measures to ensure remediation). Many of these areas explicitly reference or harmonize with existing instruments such as SE No.9/2023 and the PDP Law. SE No.9/2023.

Implementation Framework

Implementation in the draft is phased and risk-based. Lower-risk AI retains voluntary compliance plus baseline transparency obligations; high-risk AI triggers mandatory pre-deployment risk assessments, documentation, registration/notification, third-party conformity or testing (as defined by implementing rules), and enhanced human oversight requirements. The draft anticipates sectoral implementing rules to operationalize thresholds, technical standards, and enforcement mechanisms. It also recommends capacity building for regulators, accredited testing laboratories, and guidance documents for SMEs to reduce compliance burden while protecting public interests.

Monitoring and Evaluation

The draft outlines monitoring mechanisms including market surveillance, periodic reporting by PSEs, incident notification requirements, and performance indicators tied to safety, fairness and privacy outcomes. It envisions a national AI registry or dashboard to track high-risk systems, a complaint-handling channel for affected persons, and regular public reporting by Komdigi and partner agencies on compliance trends. Independent evaluation — including audits and public transparency reporting — is recommended to measure policy effectiveness and to feed iterative updates to the framework.

Penalties, Liability, and Appeals

As a concept draft, detailed sanctions are deferred to implementing regulations, but it enumerates possible measures: administrative fines, remedial orders (model rollback, algorithmic adjustments), suspension or blocking of services, revocation of PSE registration, and referral for criminal investigation where statutory offences are implicated (e.g., violations of PDP Law or ITE provisions). The draft also describes redress pathways for individuals including complaint mechanisms and requirements for transparent appeals procedures. Enforcement is envisaged to be proportional, risk-based and accompanied by guidance to support compliance.

Relationship to Other Instruments

The draft explicitly connects to: the Ministerial Circular/SE No.9/2023 on AI ethics; Law No.27 of 2022 on Personal Data Protection (PDP Law); existing Electronic Information and Transactions (ITE) law and implementing PSE rules; and sectoral laws (health, finance) where additional safeguards apply. The draft is positioned as a normative bridge that informs future binding instruments (including a proposed Presidential Regulation) while respecting existing statutory obligations under the PDP Law and other sectoral regimes. SE No.9/2023; Law No.27/2022 (PDP Law).

International Alignment

The Concept Draft emphasizes alignment with international best practices and soft-law instruments (e.g., OECD AI Principles, EU AI Act concepts), while calibrating to Indonesia’s legal context. It encourages interoperability (data protection, cross-border cooperation on AI safety and incident response) and proposes participation in multilateral fora to harmonize standards and avoid fragmentation, particularly for cross-border AI services and global platforms operating in Indonesia. This international orientation is intended to facilitate trade, investment, and cooperative enforcement mechanisms.

Implementation Timeline

Sources and References