<![CDATA[The integration of artificial intelligence into justice systems offers efficiency and access gains, but raises human rights and constitutional concerns. In Indonesia, judicial institutions have begun piloting AI-supported tools for court administration and limited decision support across civil and criminal cases, yet adjudication remains fully human. The absence of a tailored governance framework poses risks to the rule of law, judicial independence, and equality before the law. This study analyses normative weaknesses in Indonesia’s regulation of judicial AI and formulates a rights-based governance model. The method used is normative legal research employing statutory and conceptual approaches based on secondary legal materials. The results show that unclear boundaries around AI functions can turn decision-support outputs into de facto determinants of procedural direction, weakening fair-trial guarantees when parties cannot know, challenge, or obtain human review. The findings further indicate that governance must mandate human-in-the-loop control, transparency, and contestability, alongside cybersecurity, operational reliability, audit logging, and bias mitigation. Inclusive access requires hybrid service channels, so digitalisation does not exclude communities in areas with limited infrastructure. The conclusion is that Indonesia should strengthen rights-based governance so that AI improves efficiency without eroding constitutional guarantees and the legitimacy of the rule of law]]>
