<![CDATA[This study examines the construction of legal accountability for the use of Artificial Intelligence in the supervision of illegal fishing in Indonesia's Exclusive Economic Zone. The main problem of this research is not only the state's ability to detect ships, but the vacuum of norms regarding who is responsible when the results of detection, predictions, or recommendations based on algorithms are wrong, biased, unverifiable, or cause law enforcement actions that harm certain parties. This study uses normative legal methods with limited legislative, conceptual, and comparative approaches. Primary legal materials include the Fisheries Law, the EEZ, UNCLOS 1982, provisions on electronic evidence, personal data protection, and marine and fisheries resources supervision rules. Secondary legal materials were obtained from the latest journals on maritime surveillance, IUU fishing detection, remote sensing, and algorithmic accountability. The results show that Artificial Intelligence should be placed as a decision support system that requires human validation, not as an automatic authority that replaces law enforcement officials. The results of AI analysis can be valuable as intelligence information or initial indications, while its use as evidence must meet data validity, system integrity, audit trail, and chain of custody. Legal accountability is constructed in stages, i.e. the state is responsible for the policies, standards, procurement, and audit of the system; Ministries/agencies and operators are responsible for use and validation; the developer or vendor is liable for design defects, security failures, or system malfunctions; and illegal fishing perpetrators remain responsible for substantive violations. The study recommends specific norms regarding algorithmic auditing, automated decision restriction, human-in-the-loop, cybersecurity, data interoperability, and AI-based action objection mechanisms.]]>
