EU법제상 AI제조물에 대한 책임과 안전

This article examines the legal liability of AI products with a focus on the newly established Product Liability Directive and the Product Safety Regulation recently enacted by the European Parliament and the European Council. The Product Liability Directive and the Product Safety Regulation are closely intertwined to ensure the safety of products in the EU and to protect users of AI products. The Product Liability Directive expands the concept of a product to include digital manufacturing documents, AI systems, and software. The concepts of who is liable and what constitutes a product defect have also been broadened. In court, the burden of proof is on the victim to prove only the probability that the defect caused the damage. On the other hand, the Product Liability Directive is closely linked to the AI Act in terms of increasing the reliability of AI technology. The Product Safety Regulation set out minimum standards for the safety of products, which, in the context of product liability, are considered to be relevant to the target conditions of products. Under the Regulation, cybersecurity or software updates play an important role in the safety of products. In addition, online service providers have special obligations for the safety of their products in the sense that they have a unique position in the supply chain and that they have vast amounts of data on manufacturers, products and consumers. A manufacturer's compliance with the criteria in the Product Safety Regulation does not automatically lead to a product being free from defects; additional safety requirements under the Product Liability Directive should be taken into account. And product liability, product safety and market surveillance are considered as an integrated system.