On 28 September 2022, the European Commission (“EC”) adopted two new proposals which aim to adapt liability rules to the digital age with in particular a major focus on AI.
A revised Product Liability Directive
The first proposal, a revised Product Liability Directive (“PLD Proposal”), builds on known rules such as the strict liability of manufactures for unsafe products, whilst modernising their scope to encompass digital products, circular economy business models and global supply chains.
Products in the digital age: the PLD Proposal confirms that AI systems, software and AI-enabled goods are “products” within the scope of the PLD Proposal meaning that an injured person can claim compensation when a defective AI product causes death, personal injury, property damage or data loss in case specific requirements established in the PLD Proposal are fulfilled.
Circular economy: as the number of products which can (and sometimes even have to) be modified or upgraded after being placed on the market are increasing, the PLD Proposal aims to ensure that consumers have the same rights to compensation for harm caused by defective modified products as they would for entirely new products. Therefore, the rules of the PLD Proposal are designed to also apply to remanufacturers and other businesses that substantially modify products where they cause damage to a person.
Global supply chains: the PLD Proposal takes into account the vast number of products manufactured outside of the EU but in circulation in the EU market (for example, when consumers purchase products from outside the EU directly) and aims to ensure that consumers can claim compensation from an EU-based economic operator such as the manufacturer’s EU representative.
Proposed procedural changes: with the declared aim of putting consumers on equal footing with manufacturers, the PLD Proposal for the first time includes provisions that require manufactures to disclose evidence, which are supposed to ease the burden of proof on consumers.
Limitation period in case of latency injury: in order to better protect claimants and, at the same time, seek more certainty, the PLD Proposal introduces a limitation period of 15 years (instead of 10 years) in case of latency of personal injury.
Exemption from liability: the PLD Proposal will no longer give member states the possibility to derogate from the exemption afforded to manufacturers for scientifically and technically undiscoverable defects.
A brand new AI Civil Liability Directive
The second proposal, a first of its kind AI Civil Liability Directive (“AILD Proposal”), targets harmonisation of non-contractual liability rules for AI-enabled products and services. In the EC’s view, because current national rules could not properly cater for AI-based liability claims given specific characteristics of AI (complexity, autonomy and opacity, the “black box” effect), the EC proposes a harmonised EU-level staged approach for fault-based liability to prevent fragmentation among member states and increase legal certainty.
With the aim of ensuring more protection for any kind of potential victims (individuals or businesses) of any type of AI-caused damage covered by national law (e.g. life, health, property, privacy), the AILD Proposal introduces two main features:
Presumption of causality: in claims where a relevant fault has been established and a causal link to the AI performance is deemed as reasonably likely, a targeted rebuttable presumption of causality shall apply under cumulative conditions defined in Article 3 of the AILD Proposal. These conditions differ depending the AI system in question is considered a high risk system (i.e. AI creating a high risk to health, safety or fundamental rights) or not. For the EC, this presumption of causality aims at addressing challenges faced by claimants in explaining in detail how damage was caused by a certain fault or omission and establishing the causal link between such non-compliance and the output of the AI system (or lack thereof) that caused the damage.
Right of access to evidence: in cases where high-risk AI is involved, under the AILD Proposal, a court may order disclosure of relevant evidence related to the specific high-risk AI system that is suspected of having caused damage (subject to certain safeguards).
Amendment to Directive (EU) 2020/1828: last but not least, the AILD Proposal shall be added to the list set in Annex I of the EU's Collective Redress Directive as an area where consumer representative actions must be implemented by the member states.
The second stage involves a monitoring programme under which the EC will review information on incidents involving AI systems and re-assess whether more stringent measures are necessary, including whether strict liability would be more appropriate for AI cases with a particular risk profile, or whether mandatory insurance should be introduced.
It is not a coincidence that two proposals were introduced together at the same time. The EC takes the view that these two initiatives are closely linked and form a single package in its holistic approach to AI policy-making. Although claims falling within their scope deal with different types of liability – the PLD Proposal covers manufacturer’s no-fault liability for defective products, whilst the AILD Proposal aims to harmonise liability rules in fault-based claims for compensation – the EC believes they complement each other in order to form an overall effective liability system adapted to specificities of AI.
The EC also describes the AI Act (which was presented by the European Union in April 2021) and the AILD Proposal as “two sides of the same coin” which apply at different moments and reinforce each other the AI Act aims at preventing damage while the AILD lays down a safety-net for compensation in the event of damage.
Will the UK follow?
It is not clear at the time being if the UK government will follow the EU in its proposed directives and adopt similar measures to update its product liability rules addressing the use of AI systems. The UK government has however flagged AI technology as a key agenda for its national growth in the AI Regulation Policy Paper and National AI Strategy Guidance.
The Commission’s proposals will now need to be adopted by the European Parliament and the Council. Hogan Lovells is actively monitoring developments in this space - keep an eye out for our future updates.
Authored by Valerie Kenyon, Christelle Coslin, Christian Di Mauro, Gonzalo Ardila, Nicole Saurin, Lorena Baltazar, Daniel Lee, Paolo Lani and Hortense Derrien.