Research on product safety and liability for AI by Gitta Veldt and Tycho de Graaf
A recently published article, Productveiligheid en aansprakelijkheid voor AI (Product Safety and Liability for AI), by Gitta Veldt and Tycho de Graaf examines the European Commission’s proposed Artificial Intelligence Regulation.
The paper addresses the objectives and the scope of the new regulation and reviews how it relates to existing product safety law. It focuses on high risk AI systems and the consequences of the regulation for businesses, in particular, the provider and the user. The article explains what the risk-based approach in the proposal means and why high risk AI systems deserve special attention. Moreover, it discusses the consequences of the draft regulation for liability for AI systems in the legal relationship between the provider and the user and between third parties and the provider.
The authors conclude that the relationship between the draft regulation and existing product safety law is complicated because of the chosen multi-layered approach. In this approach, the draft regulation applies in addition to existing product safety law. The authors expect the scope of the draft regulation to be extended by adjusting product safety law for specific industries. This will make compliance more complex. Furthermore, the draft regulation affects the duty of care under private law. This affects manufacturers and users who are dependent on the developer of that AI system to meet their product safety obligations and have no/little influence on damage caused by the use of the product. In this situation, the authors advise these manufacturers and users to include public law product safety obligations and sanctions, as well as private law remedies in their contract.
Gitta received her PhD last year on a dissertation on product safety and liability titled 'European product standards and private law standard setting' and is currently conducting empirical legal research to market regulation and private law in the context of the awarded sector funds, in addition to traditional legal research as in the present article. De Graaf conducted his research for the article under SAILS (Society, Artificial Intelligence and Life Sciences). SAILS is a university-wide initiative aiming to facilitate collaboration across disciplines on the use of AI and is subsidized by the Executive Board of Leiden University.