The liability of robotic and autonomous systems
A comparison with other regimes
In my first and second posts, I discussed the liability of autonomous systems under UAE law and the remedies available for injured persons.
In this post, I will compare the UAE liability regime to others, in particular the European Union regime and its approach to the liability of autonomous systems. I will also discuss the approach under common law, the regime applied in England, USA, Australia and other parts of the world.
To summarise the first two posts, the UAE liability system, under the UAE Civil Code, is mainly based on the regime of “tort” or “acts causing harm” and the remedies available are:
- civil (e.g. damages); or
- criminal (e.g. fines or imprisonment).
The “product liability” regime is used to a lesser extent under the UAE consumer protection law.
A — EU Position
When we discuss EU law, we are looking mainly at a civil or continental law regime, making the provision largely similar to that in the UAE. In comparison, the Anglo-Saxon regimes in the UK, USA and other parts of the world are based mostly on common law.
The overarching principle under the domestic laws of the EU state members is “tort” law, requiring an act of harm, damage and a causality link. In some jurisdictions, causality is presumed.
These laws include rules introducing fault-based liability with a relatively broad scope of application. Sometimes accompanied by more specific rules which either:
- “modify the premises of fault-based liability (especially the distribution of the burden of proving fault); or
- establish liability that is independent of fault (usually called strict or risk-based liability).
Most liability regimes contain the notion of liability for others (often called vicarious liability)”. Liability for Artificial Intelligence and Other Emerging Digital Technologies.
However, there is no unified tort law for the EU; each member state has its own legislation. The only regulation that applies across the whole of the EU is the product liability law under Directive 85/374/EEC which came into force in 1985.
This established the principle of liability without fault (strict liability) applicable to European manufacturers. Here, where a defective product causes damage to a consumer the producer may be liable even without negligence or fault on their part.
While the existing regime has been tested in courts, when it comes to robotics and autonomous systems (RAS), gaps have been identified, especially in the case of full autonomy, where a human cannot be held liable and where defects cannot be attributed to the manufacturer.
Closing the gap
There have been many initiatives within the EU in recent years to identify such gaps and propose solutions:
- A common EU approach to liability rules and insurance for connected and autonomous vehicles — European Parliamentary Research Service, February 2018
- Liability for Artificial Intelligence and other emerging digital technologies — the Expert Group on Liability and New Technologies Formation, 2019
- White Paper on Artificial Intelligence — A European approach to excellence and trust — European Commission, February 2020
By identifying gaps in the legislation we can make sure regulation and legislation can keep up with the latest technologies and protect consumers.
The principal gap identified with the introduction of RAS is that under the current Product Liability Directive the liability of the manufacturer is mainly restricted to defects at the time of production: not suitable for an autonomous system which continues to learn and take decisions post-production.
I will discuss this in more detail in my final post and will propose solutions for the UAE.
B — Common Law
Common law applied in England and Wales, the USA, Australia and other parts of the world is a different system to the Civil Law applied in the UAE. This does not necessarily mean that the position on liability for harms caused by autonomous systems is completely different; to the contrary, there are many similarities.
However, the legal terminologies may differ. For instance, as I pointed out in my first post, “tort” as a term may not have an equivalent under the UAE Civil Code but there are provisions of “acts causing harm”, which is basically the common law tort regime. Common law is applied in the Dubai International Financial Centre and Abu Dhabi Global markets, but I don’t go into detail on these in my posts.
Liability in tort under common law is basically a fault-based system, where a party owes the other a duty of care. In the case of a breach of such duty of care, then the party is liable in tort and needs to compensate the injured person. Tort deals mainly with negligence, and it is a different regime from product (strict) or contractual liability.
Common law also covers “vicarious” liability which is usually found in the employer/employee relationship where the person who suffered harm can sue both the person who caused the damage (an employee of a company for example) and the person who was “vicariously liable” (the company itself/employer). Vicarious liability exists under the UAE Civil Code article 313.b. This relationship can be mirrored to an autonomous vehicle and the party controlling it.
What does this mean for autonomous systems?
The main challenge with common law for RAS is being able to prove liability in a tort claim for product liability cases as the damage may be too remote. Negligence is also difficult to prove. So while consumer protection laws would be more suitable for product liability claims, as generally you do not have to prove fault, as is the case in the EU regime, there is a gap when it comes to autonomous systems with common law too. A number of groups, including AAIP, are working to close such gaps.
As I pointed out in my previous post, for the UAE product liability claims are not very common, especially under consumer protection laws. The common remedy for injured persons is claims in tort. This is one of the main differences between the UAE regime (and generally regimes in the region) opposed to western regimes. However, while the discussion is at very early stages in this region, I still believe that, similarly to EU and common law, there is a liability gap for full autonomous systems and will discuss this in my next post.
Tarek Nakkach
Regional Legal Counsel — UKIMESA
Hewlett Packard Enterprise
and AAIP Programme Fellow
Contribute to the strategic development of the Programme as a Fellow.
assuring-autonomy@york.ac.uk
www.york.ac.uk/assuring-autonomy
This post is intended to be used for information purposes only and not as legal advice.
For the purpose of this post, the term autonomous systems covers robotics and Artificial Intelligence (AI) software or algorithms; any system that could function without human intervention or control. The focus from a legal point of view is on the term “autonomous”.