What do Indonesian regulations say about the role of AI and physicians in fitness-for-work assessments?

No Indonesian statute currently regulates artificial intelligence specifically for occupational fitness-for-work assessments. Instead, occupational health regulations assign medical examinations to physicians, while data protection law and AI ethics guidance restrict decisions made solely by automated machines.

Medical examinations sit with the physician. Law Number 1 of 1970 (UU 1/1970) Article 8(1) requires workplace managers to have the physical health, mental condition. Furthermore, Physical capability of prospective or transferring workers examined according to the nature of their work. Article 8(2) mandates periodic examinations by a physician appointed by the employer and approved by the Director designated by the Minister of Manpower. Permenaker PER-02/MEN/1980 Article 1 letters a to c defines pre-employment, periodic, and special examinations as examinations conducted by a physician. In addition, Article 2(1) states that pre-employment examinations verify that the worker is suited to the assigned job. PP 28/2024 Article 233(5) letters d and e likewise list medical examinations and fitness-for-work assessments as preventive occupational health measures.

Assessment guidelines are also written by humans. Permenaker PER-02/MEN/1980 Article 2(5) requires the employer or manager together with the physician to draw up pre-employment examination guidelines approved by the Director. In addition, Article 3(4) requires the same for periodic examinations. In our interpretation, any rule set used by software to prepare a draft fitness category should trace back to those physician-approved guidelines.

Two instruments address AI and automated decisions. First, SE Menkominfo 9/2023 item 6(c)(2)(b)—an ethical guideline rather than a statute—states that AI should not serve as a policy determinant or decision-maker affecting humanity. Second, Law Number 27 of 2022 (UU 27/2022) Article 10(1) protects data subjects. They have the right to object to automated decisions producing legal effects or significant impact. Article 34(2) letters a and b classify significant automated decisions and processing of specific personal data. Which includes health data under Article 4(2)(a)—as high-risk processing requiring a Data Protection Impact Assessment (DPIA). Because a fitness-for-work conclusion affects job placement. Keeping the final determination with the examining physician while AI only prepares a draft recommendation means the outcome is not based solely on automated processing.

Examination results are further bound by medical confidentiality under Law Number 17 of 2023 (UU 17/2023) Article 4(1)(i) and Article 301(1). Because Article 4(4) exceptions—such as patient request (letter f) and statutory mandates (letter h). Do not automatically grant employers raw medical access, individual clinical reports stay with the patient and medical team while employers receive the fitness-for-work conclusion.