Comments on the article titled “The regulation of AI liability in Europe: a critical overview of two recent Directive Proposals – the (new) ALD and the (revised) PLD” by Beatriz Garcia

Abstract

This commentary expands on Beatriz Garcia’s critique of the Artificial Intelligence Liability Directive (AILD) and the revised Product Liability Directive (PLD), focusing on their interaction with the recently enacted AI Act. Garcia identifies key issues: the PLD’s narrow focus on software as a “product” and the AILD’s burdensome evidentiary requirements for claimants seeking AI-related compensation. The AI Act, effective August 2024, establishes a risk-based regulatory framework prioritizing harm prevention over liability. While fostering innovation through regulatory sandboxes, it lacks civil compensation mechanisms, creating regulatory gaps. Case studies—an autonomous vehicle fatality and discriminatory AI hiring practices—highlight the difficulties of applying AILD and PLD frameworks, particularly in addressing non-physical harm and the “black box” nature of AI systems. This analysis underscores a disconnect between the AI Act’s harm-prevention focus and the liability frameworks, favoring companies while complicating compensation for individuals. A recalibrated regulatory approach is needed to align innovation with accountability.