The New Product Liability Directive and Innovations in AI

On November 18, 2024, Directive (EU) 2024/2853 was published in the Official Journal of the European Union, repealing the previous Directive 85/374/EEC of July 25, 1985, effective from December 9, 2026.  

After nearly 40 years, the product liability framework has been modernized with important changes.  

Directive 85/374/EEC, implemented in Italy through Presidential Decree No. 224/1988 and later incorporated into Legislative Decree No. 206/2005 (the “Consumer Code”), has ensured a high level of consumer protection for almost four decades.  

However, the arise of new technologies —particularly the increasing uptake of software in marketed products and advancements in artificial intelligence (“AI”)— demanded a revision of the existing regulations.  

The directive clarifies from its first recitals that liability for defective products must be strict in order to ensure fair risk distribution in modern technological production.  

The European legislator acknowledged the need to revise the repealed Directive 85/374/EEC becauseof recent technological developments, such as the proliferation of AI systems. This led to an update in the definition of "product" within the new Product Liability Directive (PLD) that now explicitly includessoftware (and thereby AI systems) as intangible products of the digital era.  

The primary goal of this revision is twofold:  

First, the directive refers to AI Act, establishing that software producers or developers, including providers of AI systems under Regulation (EU) 2024/1689 (AI Act), are considered “manufacturers” for the purposes of defective product liability.  

Another key aspect concerns substantial modifications to software or AI systems.  

If an update, improvement, or continuous learning process in an AI system constitutes a substantial modification, the modified product is treated as newly placed on the market at the time of the change (cf. Recital 40, PLD; Article 25, AI Act).  

Additionally, the definition of “component” in the PLD aligns with and is particularly relevant to high-risk AI systems under Article 6 of the AI Act.  

The term “component,” defined as “any tangible or intangible item, raw material, or related service integrated into or interconnected with a product,” is crucial for identifying high-risk AI systems under Article 6(1) of the AI Act, i.e., systems intended to be used as safety components of a product and subject to third-party conformity assessments.  

On the other hand, the AI Act already acknowledged the possibility of compensation for damages caused by AI in accordance with Directive 85/374/EEC (cf. Recital 9, AI Act).  

Causal Link Between Damage and Defect in AI Systems  ​

Furthermore, a central issue is proving the causal link between the damage suffered and the product defect, particularly in AI systems.  

The European legislator highlights the challenges faced by claimants in AI-related cases, where proving causation can be complex. In such instances, in light of the new PLD, claimants are not required to explain technical details or specific characteristics of the AI system as they might hinder the establishing of causality.  

These provisions aim to balance consumer protection and manufacturer liability, providing a regulatory framework suited to the challenges of the digital age.  

The new directive preserves the fundamental principles of previous pieces of regulations but introduces significant updates to address the needs of the current digital economy. The main updates are as follows:  

The definition of "product" has been expanded to include any movable good, even if integrated into another movable or immovable good (or interconnected with it), as well as electricity, digital manufacturing files, and software.  

This update, as mentioned above, ensures that goods incorporating software (including AI-based products) are clearly included within the meaning of "product," allowing consumers to claim compensation for damages caused by these systems.  

One of the most significant changes concerns the burden of proof on the claimant.  

Under the traditional product liability framework, claimants had to prove the product defect, the damage suffered, and the causal link between the defect and the damage.  

The new directive introduces specific circumstances where a presumption of defectiveness applies.  

This change aims to facilitate compensation claims for consumers, reducing evidentiary challenges.  

Upon request by a judicial authority, whether the claim for damages has adequate grounds, economic operators involved in the claim must provide relevant evidence within their possession. National laws must include specific measures to protect confidential information and trade secrets.  

Failure to comply with disclosure obligations will result in a presumption of product defectiveness.  

Directive (EU) 2024/2853 extends liability to logistics service providers and online platforms, thereby including e-commerce platforms in the supply chain.  

This broadens the scope of parties traditionally held liable to consumers, enhancing consumer protection in the digital age.  

The updates introduced by Directive (EU) 2024/2853 will significantly impact economic operators.  

The enhanced consumer protections will require operators to adapt to new responsibilities and obligations.  

To grasp new opportunities offered by digitalization while mitigating associated liability risks, operators must act promptly, focusing on compliance in documentation and contracts to ensure full adherence to the new provisions ahead of their implementation on December 9, 2026.