| Abstract [eng] |
Background: The question of who bears liability when an AI system causes harm has long been debated in the field of AI and law. When the EU proposed the Product Liability Directive and the AI Liability Directive in 2022, many expected the two instruments to jointly determine how such cases would be handled. This picture changed once the AI Liability Directive was withdrawn in February 2025 and the revised Product Liability Directive was adopted in October 2024. The result is that the EU now provides harmonised liability rules only for situations in which AI products injure consumers or other natural persons. With this shift, current discussions on AI liability can either examine the consequences of withdrawing the AI Liability Directive or look closely at how the updated Product Liability Directive (EU) 2024/2853 (“Directive”) allocates responsibility. This article takes the latter approach. For the consumer and naturalperson contexts to which the Directive applies, it argues that the debate should turn to analysing the specific liability structure the Directive sets out. Methods: The revised Directive is not an easy instrument to interpret, given its distinctive terminology, high level of detail, and extensive lex specialis provisions when compared to the general rule of fault-based liability. Although it replaces the 1985 Directive–and thus builds on an existing body of scholarship–the 2024 revision introduces changes that necessitate renewed analysis. The most suitable method for understanding this revised liability regime, clarifying its ambiguities, novelties, and problematic aspects, is the doctrinal legal approach. Accordingly, this article employs doctrinal analysis to examine and systematise the revised Directive. Results and conclusions: The article organises the Directive’s provisions into four categories: first, the scope of application, defining when product liability applies; second, the elements of liability, specifying what the claimant must prove; third, the defences, indicating the exemptions from liability on which the defendant may rely; and finally, the procedural rules, governing disclosure of evidence, relevant to both parties, and the conditions under which the burden of proof, resting on the claimant, may shift to address evidentiary challenges. On the basis of this fourth feature—procedural rules that ease the claimant’s evidentiary burden—the article argues that the Directive alters how EU product liability should be conceptually defined. Under the 1985 Directive, the framework rested on two conceptual axes defining strict product liability: first, replacing fault with product defectiveness; and second, limiting defences, both of which made product liability “strict”. Since the new procedural rules further strengthen the claimant’s position, the article concludes that the revised Directive adds a third axis to the concept of strict product liability, thereby making the regime even stricter through burden-alleviation rules. |
