The proposals of EU Directive on the liability of artificial intelligence and the digital age might have a considerable impact in boosting innovation within the European Union.
On September 28, 2022, the European Commission announced adopting two regulatory proposals to adapt liability rules to the digital age, the circular economy, and the impact of global value chains. Specifically, the first proposal aims to update the regime of the current Product Liability Directive, while the second proposal consists of a new directive on liability for artificial intelligence and is intended to make it easier for those who have suffered damage resulting from the use of artificial intelligence systems to obtain compensation.
The adoption of the two proposed directives follows the public consultation that closed on January 10, 2022, and is part of a coordinated European approach to address the growing deployment of artificial intelligence and digital technologies. The Artificial Intelligence Regulation proposed by the Commission in April 2021 focuses primarily on monitoring and damage prevention, while the AI Directive aims to harmonize the liability regime in cases where the artificial intelligence systems deployed exhibit traits of particular complexity and sometimes opacity.
The changes introduced by the Liability Directive aim to update and strengthen the current rules on strict liability and compensation for defective product damage, in particular:
- ensuring that liability rules are clear for companies that make substantial changes to their products;
- allowing compensation for damages when technologies, such as robots or drones, are made unsafe by software updates and mismanagement of cybersecurity vulnerabilities;
- creating a level playing field between EU and non-EU manufacturers and allowing consumers to seek redress from importers of products in the EU as an additional remedy; and
- putting consumers on an equal footing with producers through the ability to require producers to make evidence accessible, with increased flexibility in time limits for claims and easing the burden of proof for injured parties in complex cases.
As for the Artificial Intelligence Directive, the new framework aims to establish a uniform set of rules for access to information and to ease the burden of proof in cases of damage caused by IA systems. In this way, the AI Directive makes it easier for injured parties to prove the causal link between the event and damage. In particular, two solutions are adopted:
- a presumption of causation to handle injured parties’ difficulties in proving causation (particularly in the case of complex AI systems, which are often seen as “black boxes”); and
- a right of access to evidence from companies and suppliers where high-risk artificial intelligence systems are involved.
The European Commission made clear that the two proposed directives are designed to be applied in a complementary way: the updated Liability Directive will apply to claims for damages caused by defective products, including damage to health, property, and data, while the new Artificial Intelligence Directive will cover cases where the damage was caused by an artificial intelligence system. The main common feature of the two proposals is the introduction of tools to ease the burden of proof in favor of injured parties.
With respect to the relationship between the Artificial Intelligence Directive and the proposed AI Regulation, the two pieces of legislation are “two sides of the same coin” and apply to different cases within an overall framework. On the one hand, the AI Regulation introduces rules on safety in the development and use of AI systems, aimed at reducing the overall risk to stakeholders and preventing possible harm. On the other hand, the IA Directive takes action when risks that cannot be eliminated cause harm to people or property.
The AI Directive and the AI Regulation use the same definitions, first of all, maintaining the distinction between high-risk and non-high-risk artificial intelligence and recognizing the documentation and transparency requirements of the AI Regulation. The end result is a uniform approach to safety and regulation.
The Commission’s proposals will now have to be reviewed and adopted by the European Parliament and the Council.
On a similar issue, you can read the article “Is the draft EU Artificial Intelligence Regulation on the right track?“.