European Commission Proposes New Liability Rules on Products and AI
On September 28, 2022, the European Commission published two proposals—the Revised Product Liability Directive and the AI Liability Directive—aimed at adapting liability rules to the green and digital transition within the European Union ("EU").
The proposed Revised Product Liability Directive is set to modernize the current EU framework on manufacturers' liability for defective products. It will apply to all products sold in the EU, from furniture to software updates. In particular, the proposal seeks to:
- Adapt liability rules to circular economy business models, in line with the EU's efforts to achieve sustainability and waste reduction goals, by clarifying when economic operators can be held liable when substantially modifying or upgrading products.
- Adapt liability rules to the reality of products in the digital age, by allowing compensation for damage caused by, e.g., AI systems, digital services, or cybersecurity vulnerabilities.
- Create a level playing field between EU and non-EU manufacturers, by ensuring that an EU-based business can always be held liable for damages caused by products manufactured outside the EU.
- Facilitate claims to compensation, by requiring manufacturers to disclose necessary information in court and by easing the burden of proof for victims in more complex cases, such as those involving pharmaceuticals, smart products, or AI-enabled products.
The proposed AI Liability Directive provides for a targeted harmonization of national civil liability rules for AI aimed at increasing legal certainty and fostering AI innovation. It will apply to non-contractual civil law claims for damages caused by all types of AI systems. It supplements the rules under the above-described proposed Revised Product Liability Directive by introducing two main additional measures specifically for AI:
- Alleviating victims' burden of proof through the "presumption of causality," whereby courts can establish the causal link between the damage and non-compliance of providers of AI systems with a certain obligation relevant to the harm (e.g., with a duty of care under EU or national law), if the victims can demonstrate such non-compliance. Such presumption is rebuttable by proving that a different cause provoked the damage.
- Empowering courts to order providers of high-risk AI systems (as defined under the proposed AI Act) to disclose relevant information, subject to appropriate safeguards to preserve the legitimate interests of all parties, such as trade secrets or other sensitive information.
The Commission's proposals will now be examined by the European Parliament and the Council of the EU, in parallel with the above-referred proposed AI Act.
