Civil Liability for Artificial Intelligence–Related Harm between Contractual and Strict Liability (2021-2025): A Comparative Analytical Study of Selected Arab Legal Systems and the European Union

Abstract

This article examines civil liability for harm caused by artificial intelligence (AI) amid its rapid deployment in several domains. AI systems-often operating through deep learning and quasi-autonomous decision-making-raise evidentiary challenges associated with the “black box” problem, making it difficult to prove fault and causation under traditional liability models.
The article distinguishes between two situations: (1) AI operating under effective human control, where contractual liability may remain workable in cases of breach of contractual obligations; and (2) AI acting with high autonomy, where strict (objective) liability becomes more appropriate to ensure effective compensation without requiring proof of a clearly identifiable human fault.
Methodologically, the article adopts an analytical and comparative approach, critically comparing the Arab legal framework (Kuwait, Egypt, and Jordan-which largely relies on general rules not designed for complex digital harms-with the European model, which has recently advanced regulatory and compensatory mechanisms (notably risk-based regulation and eased evidentiary burdens). 
The article concludes by proposing a modern framework-particularly for Kuwait-grounded in risk allocation, effective control, and complementary mechanisms such as mandatory insurance and compensation funds.