Decree of the Minister of Manpower Number 103 of 2026 Establishing Generative AI Competency Standards and Requiring Adjustments to Competency Certification Within 6 Months

Introduction

On 6 April 2026, the Minister of Manpower issued Decree of the Minister of Manpower Number 103 of 2026 on the Establishment of the Indonesian National Work Competency Standards for the Category of Telecommunications, Computer Programming, Consulting, Computing Infrastructure, and Other Information Services Activities, Main Group of Programming, Computer Consulting, and Related Activities in the Area of Expertise of Artificial Intelligence, Sub-area of Knowledge-Based Systems (“MoM Decree 103/2026”). MoM Decree 103/2026 establishes and updates the standards used in structuring national qualification levels and implementing education, training, and competency certification in the field of Artificial Intelligence ("AI").

MoM Decree 103/2026 was drafted to adapt the Indonesian National Work Competency Standards (Standar Kompetensi Kerja Nasional Indonesia, "SKKNI") to technological developments, particularly in the Knowledge-Based System sub-area. In its recitals, the establishment of MoM Decree 103/2026 follows up on the results of the national convention on 21 October 2025 and the proposal from the Head of the Center for the Development of Communication and Digital Human Resources Ecosystem, Ministry of Communication and Digital. MoM Decree 103/2026 sets competency standards to support the availability of a workforce capable of designing, developing, and managing artificial intelligence solutions according to industry needs, including in the aspects of data protection and system reliability.

Comparison

MoM Decree 103/2026 repeals the provisions of and invalidates Decree of the Minister of Manpower Number 123 of 2021 on the Establishment of the Indonesian National Work Competency Standards for the Category of Information and Communication, Main Group of Programming, Computer Consulting, and Related Activities (YBDI), Area of Expertise of Artificial Intelligence, Sub-area of Knowledge-Based Systems (“MoM Decree 123/2021”). The comparison between MoM Decree 103/2026 and MoM Decree 123/2021 is as follows:

Key Provisions

Legal Standing and Qualification Standard Establishment

The SECOND Dictum establishes the SKKNI document as a reference for structuring national qualification levels and implementing education, training, and competency certification. The THIRD Dictum stipulates that the implementation of the SKKNI and the structuring of national qualification levels are established by the Minister of Communication and Digital and/or related technical ministries/agencies in accordance with their duties and functions. These provisions apply to technology service provider businesses and educational institutions in formulating training programs and workforce absorption.

Planning and Design of AI Solution Architecture

The Annex section (Unit Codes K.62AIN00.001.2, K.62AIN00.002.2, K.62AIN00.003.2, and K.62AIN00.004.2) stipulates that AI practitioners must understand the planning stages prior to system development. These stages include the formulation of business objectives, the determination of performance metrics, and the preparation of technical architecture. The workforce must prepare a project plan covering risk identification, data requirements, and estimated development costs.

Development and Validation of Expert-Based Models (Knowledge-Based System)

The Annex section (Unit Codes K.62AIN00.010.2, K.62AIN00.011.2, K.62AIN00.012.2, and K.62AIN00.013.2) governs workforce competency in the development of artificial intelligence-based expert systems. AI practitioners must design knowledge representation schemes (such as knowledge graphs or semantic networks), gather knowledge from expert sources or literature through observations and interviews, and structure it into the system's knowledge base. Furthermore, these provisions cover structural verification and domain-based knowledge validation to ensure consistency and feasibility before the AI model is utilized.

Data Management and Machine Learning

The Annex section (Unit Codes K.62AIN00.010.2 to K.62AIN00.013.2) governs competencies related to dataset management. AI practitioners must understand data labeling procedures, assess data quality, and conduct filtering or feature engineering, including data processing to reduce bias or anomalies. These provisions direct companies to implement data preprocessing stages before training algorithms.

Technical Standardization of Data Labeling and Reconstruction

The Annex section (Unit Codes K.62AIN00.010.2, K.62AIN00.012.2, and K.62AIN00.013.2) governs data management during the preprocessing stage. Companies must formulate and implement data labeling or annotation procedures. The workforce must understand the data processing process, including feature engineering, data transformation such as normalization and scaling, and data quantity adjustments through oversampling or undersampling during the data filtering stage, to reduce data imbalances that may affect processing results.

Mastery of the Generative Model Ecosystem (Generative AI)

The Annex section (Unit Codes K.62AIN00.014.1, K.62AIN00.015.1, K.62AIN00.016.1, K.62AIN00.017.1, and K.62AIN00.018.1) governs competencies related to the use of generative models. The workforce must validate AI system outputs, which includes the following:

1. Formulating appropriate instructions (prompt engineering);

2. Conducting model adaptation (such as fine-tuning or transfer learning) using relevant data;

3. Implementing knowledge augmentation techniques (such as Retrieval Augmented Generation/RAG) to mitigate the risk of inappropriate outputs; and

4. Evaluating generative model results to assess consistency, relevance, and potential bias.

Implementation of Ethics, Reliability, and User Privacy Protection

The Annex section (Unit Codes K.62AIN00.019.1, K.62AIN00.020.1, K.62AIN00.021.1, K.62AIN00.022.1, and K.62AIN00.023.1) governs the implementation of ethical principles in the use of AI. The workforce must implement the principle of fairness, ensure system results are explainable (explainability), and maintain system reliability. Under Unit Code K.62AIN00.022.1, AI developers must implement privacy protection through techniques such as anonymization and encryption, as well as prepare accountability mechanisms in the event of data breach or violation incidents.

Implementation of Algorithm Explainability (Explainable AI/XAI)

The Annex section (Unit Code K.62AIN00.020.1) governs the implementation of explainability in AI solutions. This provision directs the use of explainable AI systems, including in the decision-making process. AI practitioners must use explanatory tools (such as SHAP, LIME, or attention visualization) to present analysis results in a format understandable to both users and internal parties. This implementation aims to support transparency and aid in the assessment of the results generated by the system.

Assurance of Accountability and Periodic Audit Obligations

The Annex section (Unit Code K.62AIN00.023.1) governs the implementation of accountability in the development and use of AI. The workforce must formulate accountability objectives from the design stage and develop accountability indicators based on accountability dimensions, which include context, range, actors, forums, standards, processes, and implications. This provision requires that accountability implementation be documented and periodically evaluated to identify accountability gaps.

System Integration, Deployment, and Maintenance

The Annex section (Unit Codes K.62AIN00.024.2, K.62AIN00.025.2, K.62AIN00.026.2, and K.62AIN00.027.2) governs the final stages of AI development. The workforce must integrate technical architecture components into the operational environment (deployment). Additionally, the workforce must prepare a maintenance plan, continuously monitor system performance post-deployment, and conduct internal testing with relevant parties to support system operations.

Transitional Provisions

The FIFTH Dictum stipulates that the implementation of competency standards must be adjusted to the latest SKKNI within a maximum of 6 (six) months from the date of establishment (6 April 2026). In addition, the FOURTH Dictum stipulates that these competency standards are subject to periodic review every 5 (five) years or as deemed necessary.

Closing

MoM Decree 103/2026 establishes competency standards in the field of artificial intelligence covering all stages of system development and implementation, from planning and design, data management and machine learning, and knowledge-based model development, to the use of generative models, as well as the implementation of ethical principles, system reliability, privacy protection, and accountability. These standards are used in structuring national qualification levels and implementing education, training, and competency certification, thus impacting the adjustment of training materials, certification processes, and workforce competency requirements in the AI field. All relevant parties must adjust their implementation of these competency standards no later than 6 (six) months from 6 April 2026. Therefore, companies, training institutions, and relevant parties must review the readiness of their programs and workforce to ensure compliance with the applicable standards.