Pertanggungjawaban Pidana Atas Kejahatan Berbasis Algoritma: Evaluasi Model Hallevy dalam Konteks Pembaruan Hukum Pidana Indonesia

Danto Herdianto (Sekolah Tinggi Hukum Galunggung)

Jurnal Hukum Lex Generalis
Vol 7 No 7 (2026): Tema Hukum Pidana

The development of AI raises criminal attribution problems when algorithmic outputs produce consequences that satisfy criminal elements. This normative juridical study examines Gabriel Hallevy’s three models and develops an AI Criminal Liability Matrix using AI autonomy, human fault, control, causation and applicable legal norms. A hierarchy analysis shows that Indonesian law does not specifically establish the Matrix as a legal instrument or AI as an autonomous criminal subject. The resulting condition is a regulatory gap rather than a total legal vacuum. Comparison with the European Union and South Korea highlights risk-based governance, human oversight, documentation and actor-specific obligations. These findings emphasize the need to strengthen a criminal liability framework capable of identifying and attributing fault to actors involved in the development, deployment and use of AI.