Isabel Foley and Domhnall Breatnach, Arthur Cox
New and emerging technologies – such as autonomous, self-learning and artificially intelligent products, advanced robotics and the Internet of Things – are an increasing feature across a variety of sectors, including transport, healthcare, agriculture, IT, data management and communications. Such technologies offer a range of societal and economic benefits, from prevention of human error and failure, and improved safety, to increased productivity and economic growth. Although the existing EU and Irish product liability rules, which have been in place for many decades, have proven themselves reasonably flexible in adapting to the legal challenges faced by technological advancements in the pre- and early-digital age, new and emerging technologies present novel and more complex legal challenges. These new and emerging technologies raise a series of fundamental questions. Who is, and who should be, liable when they cause harm? Can the existing product liability rules be suitably adapted to accommodate these technologies and, if so, how?
The Liability for Defective Products Act 1991 (LDPA) implemented EU Directive 85/374/EEC on liability for defective products (the Product Liability Directive) into Irish law. The Product Liability Directive was adopted in 1985 – the same year the first .com Internet domain name (symbolics.com) was registered and Microsoft’s Windows operating system was first released. Technologies have evolved dramatically since. In 1985, European lawmakers were clear in the view (as recorded in the recitals to the Product Liability Directive) that “liability without fault on the part of the producer is the sole means of adequately solving the problem, peculiar to our age of increasing technicality, of a fair apportionment of the risks inherent in modern technological production.” This core principle of no-fault producer liability was expressed in simple terms in article 1 of the Product Liability Directive: “The producer shall be liable for damage caused by a defect in his product.”
The product liability regime laid down by the Product Liability Directive may represent a reasonable and fair apportionment of risks in the context of traditional products, where producers exercise largely complete control over the design and manufacture of their products and can readily assess and foresee their risks. However, by comparison, new and emerging technologies present an array of additional challenges and risks for producers, which necessitate a reassessment of the existing liability regime. For example, in the case of autonomous, self-learning and artificially intelligent products, the level of input that the product may require from a user, and the amount of control that a user may have over the product, can vary dramatically – from low or partially automated products to high or fully automated ones. Such varying levels of automation complicate matters. A problem with a product could conceivably arise as a result of: a defect in a hardware or software component; a deliberate choice of the autonomous product; interaction between the user and the product; a self-learned choice based on prior user patterns or behaviours or other extrinsic factors; or receipt of incorrect data or information from an external network outside both the manufacturer’s and the user’s control. Who should bear responsibility in such circumstances?
Whether existing liability rules can be adequately adapted, or, indeed, a single set of effective rules can be designed, to cover the ever-expanding array of new and emerging technologies remains to be seen. These technologies raise questions about all of the core legal components of the existing product liability regime, including what we understand as a “product” and a “defect”. Under the Product Liability Directive, and the LDPA, a “product” is defined in terms of “movables”. Although few difficulties arise in understanding this in the context of traditional products, the position is not so clear when it comes to new and emerging technologies. Some new and emerging products are, in fact, complex integrated and interconnected systems, comprised of not only various hardware components, but also software elements (embedded and non-embedded), which, in turn, may rely on critical data inputs from a variety of networks. Is the “product”, in reality, multiple products? Should the producer of each component be equally liable for any damage caused by the product as a whole?
Further questions arise when considering what is meant by a “defect”. Under the Product Liability Directive and the LPDA, a product is currently considered defective if it fails to provide the safety which a person is entitled to expect, taking all circumstances into account – including the presentation of the product, the use to which it could reasonably be expected that the product would be put, and the time when the product was put into circulation. However, in the case of new and emerging technologies, what level of safety should a person be entitled to expect, for example, from an autonomous, self-learning or artificially intelligent product, where the product’s ultimate performance may be dependent on learning based on user input, prior usage or behaviour patterns, or extrinsic factors or data sources?
The legal challenges posed by new and emerging technologies are not just limited to the statutory defective products liability regime. They also arise in the context of common law contract and tort principles. Under the general common law principle of duty of care that applies in Ireland, a product manufacturer owes a duty of care to all those who may be foreseeably injured or damaged by their products. Although these principles of negligence established in the seminal case of Donoghue v Stephenson have stood the test of time, and have proven themselves flexible enough to apply not just to the dead snail in Mrs Donoghue’s bottle of ginger beer, but to much more sophisticated products that have followed, it remains to be seen whether or not the principles can be adequately adapted for new and emerging technologies. As to the issues of foreseeability and causation, to what extent can the manufacturer of an autonomous, self-learning or artificially intelligent product reasonably foresee every way in which the product might ultimately function? What is, or should be, the appropriate standard of care in the context of such technologies? If a system or product reaches an appropriate standard of performance and safety during manufacturer testing, will that constitute reasonable care on the part of the manufacturer?
In Ireland, under the Sale of Goods Act 1893, as amended by the Sale of Goods and Supply of Services Act 1980, there is also an implied condition that goods are of “merchantable quality”, requiring a product to be fit for the purpose or purposes for which a product of that kind is commonly bought, and to be as durable as it is reasonable to expect having regard to any description applied to it, the price (if relevant) and all other relevant circumstances. In addition there are criminal sanctions for producers who place a product that is not a “safe product” on the market pursuant to the European Communities (General Product Safety) Regulations 2004 (GPSR), which implement Directive 2001/95/EC on general product safety (which is eventually to be repealed once the EU proposal on consumer product safety is finalised and comes into force). Under the GPSR, a “safe product” is defined as a product that, under normal or reasonably foreseeable conditions of use (including duration, and, where applicable, putting into service, installation and maintenance requirements) does not present any risk or only the minimum risks compatible with the product’s use, considered to be acceptable and consistent with a high level of protection for the safety and health of persons. Again, autonomous, self-learning and artificially intelligent products will undoubtedly present challenges when it comes to assessing if the product is fit for purpose or what the product’s “normal and reasonably foreseeable conditions of use” are.
The EU Commission is currently evaluating whether the Product Liability Directive is fit for purpose, particularly in the context of new and emerging technologies. The evaluation is to assess the Directive according to five criteria:
As part of the evaluation, the Commission launched a public consultation in 2017 to collect stakeholders’ feedback on the application and performance of the Directive, including in relation to the challenges raised by new and emerging technologies. The EU Commission’s Directorates-General for Internal Market, Industry, Entrepreneurship and SMEs; Justice and Consumers; and Communications, Networks, Content and Technology are also in the process forming an expert group to consider the issues of product liability and new technologies. The group will operate in two different groupings: a Product Liability Directive grouping and a new technologies grouping.
The Product Liability Directive grouping will consist of members appointed in a personal capacity and to represent common interests, together with members from EU-level umbrella organisations – including industry associations and federations (manufacturers, original equipment manufacturers, suppliers, retailers and repair and maintenance providers); non-governmental and consumer organisations; and academia and other research organisations and institutions – as well as from EU member-state authorities and other public entities. It will provide advice and expertise to the Commission on the Product Liability Directive and draw up guidance in relation to its implementation. The expert grouping will also be tasked with assessing the extent to which the provisions of the Product Liability Directive are adequate to solve questions of liability in the context of new and emerging technologies in light of the jurisprudence of the Court of Justice of the EU and national courts.
The new technologies grouping will assess whether and to what extent existing liability schemes can be adapted to deal with new and emerging technologies. It will also consider the issue of assignment of liability, the nature of liability (fault or non-fault based), whether it is necessary for the injured party to establish a defect, who should bear the burden of proof, and what redress possibilities insurance providers should have to recover compensated damage. The grouping will assess shortcomings in the existing liability regime and whether it is adequate to facilitate the uptake of new and emerging technologies. If not, the grouping will provide recommendations on how the existing regime should be adapted. Issues of national tort law, and any other specific national liability regimes that may be relevant, will be considered by the grouping, and questions of liability will be analysed from the perspectives of various players. The new technologies grouping will also assist the Commission with developing EU-wide principles as regards new and emerging technologies. The group’s work is be completed by June 2019.
It ultimately remains to be seen how EU policy and lawmakers (and, in the meantime, national lawmakers and the courts) will confront the many legal challenges that new and emerging technologies present, and how the issues of producer responsibility and liability will be balanced against product safety and consumer protection concerns. Such concerns are at the heart of the current product liability regime, and will undoubtedly remain so under any new or adapted one. It is important that, in seeking to apportion the risks, any new or adapted product liability regime strikes a fair balance between those competing concerns, so that the development and evolution of new and emerging technologies is not unduly impeded.