Compensatory Guarantees for Damages Caused by Artificial Intelligence in Law: A Comparative Analytical Study

Abstract

AI systems play a more significant role in decision-making in healthcare, transport, finance, and commerce. Their autonomous or adaptive operations can inflict bodily injury, cause economic loss, violate privacy, and damage reputation, while obscuring fault, causation, and the identity of the responsible actor. This paper assesses whether the conventional civil-liability rules provide adequate compensation in situations where harm results from the­ conduct of the AI provider, producer, programmer, deployer, operator, data supplier.  Through a comparative analytic method, the study looks at fault-based doctrines, presumed doctrines, strict doctrines and product liability doctrines of the civil law system.  It studies Regulation (EU) 2024/1689 on artificial intelligence, Directive (EU) 2024/2853 on liability for defective products and Article 82 of Regulation (EU) 2016/679 on compensation for unlawful personal-data processing. Next it assesses judicial damages, corrective remedies, mandatory insurance, compensation funds, joint liability and presumptions of proof. The study finds fault-based liability remains useful where negligent design, testing, deployment, supervision or use can be proved but may fail when technical opacity blocks access to evidence. As such, effective victim protection should rely on a layered model, it would combine, ordinary civil liability encompassing fault-based damage claims and risk-based duties, disclosure and record-preservation obligations, rebuttable presumptions, compulsory insurance for defined high-risk systems, and the residual compensation fund. This model enhances access to full and timely reparation while preserving legal certainty.