New liability rules for products and artificial intelligence

On September 28, the European Commission adopted the following proposals:

• Proposal: Revision of the Product Liability Directive
• The revision of the Product Liability Directive aims to modernize and strengthen the strict product liability rules on compensation for personal injury, property damage or losses caused by unsafe products. Specifically, the following changes are proposed:
• Rules for companies that substantially change their products are strengthened.
• The rules are modernized to include liability for products and systems of the digital age and to cover insecure software updates, artificial intelligence (“AI”) or digital services, and cybersecurity vulnerabilities.
• Consumers may choose to submit complaints to the importer or the manufacturer's representative in the European Union (“EU”) for the purpose of claiming compensation, with the aim of creating fairer competition conditions between EU and non-EU manufacturers.
• The directive establishes a requirement for manufacturers to disclose evidence, and it provides for greater flexibility in the time periods for bringing actions for damages. In certain complex cases, such as pharmaceutical products or AI, victims’ burden of proof is lowered.

• Proposal for a directive on adapting non-contractual civil liability rules to AI
• For the first time, the new AI Liability Directive proposes the harmonization of national AI liability rules, making it easier for victims of AI-related damage to obtain compensation while ensuring a balance between consumer protection and the AI sector:
• Rules on access to information are standardized.
• The burden of proof for damage caused by AI systems is lowered, broadening the protection for the respective victims (whether individuals or companies).
• Cases that would not be covered by the Product Liability Directive will now be covered, such as harm caused by unlawful conduct (e.g., privacy breaches and discrimination caused by processes involving AI technology).
• A "presumption of causation" is established to simplify the process of proving that the act caused the damage if it is reasonably likely that there is a causal link between the AI-related action and the damage.
• In high-risk AI cases, victims have more remedies, as they now have the right to access the evidence that companies and suppliers must file.
• Guarantees are also provided for the AI industry, including the right to contest actions for damages when they are based on a presumption of causation.
• Victims harmed by AI products or services may benefit from the same level of protection they would be entitled to if this harm resulted from any other circumstances, in line with the context already outlined in the White Paper on Artificial Intelligence and the Commission's 2021 proposal for the Artificial Intelligence Act.

The above proposals are now pending to be adopted by the European Parliament and the Council.