European Union: Published results of complementary impact assessment on the Proposal for a Directive on adapting non-contractual civil liability rules to Artificial Intelligence

On 19 September 2024, the European Parliament published the results of the complementary impact assessment on the Proposal for a Directive on adapting non-contractual civil liability rules to Artificial Intelligence (AI). The proposal aims to adapt non-contractual civil liability rules to better address the challenges posed by AI. Furthermore, it recommends extending the scope of liability to cover general-purpose and high-impact AI systems, including not just AI-specific technologies but also software. In addition, the proposal is the introduction of a mixed liability framework that combines both fault-based and strict liability, depending on the type and usage of AI systems. This approach seeks to ensure accountability across a range of AI applications, addressing the risks and uncertainties associated with emerging technologies. One of the core recommendations is to move away from an AI-focused directive and transition toward a broader regulation that encompasses software liability in general. This shift would help prevent market fragmentation and ensure clearer and more uniform legal standards across the European Union (EU). The study emphasises that such a move would provide legal clarity and coherence, which is essential for the EU’s digital single market and for fostering innovation without compromising consumer protection. Moreover, the proposal critiques the European Commission’s original impact assessment, highlighting its failure to explore a wider range of regulatory policy options. In particular, it points out that the initial assessment did not fully consider the potential benefits and drawbacks of implementing a strict liability regime for AI. Finally, the cost-benefit analysis in the original proposal is viewed as incomplete, particularly in its examination of the strict liability approach.