The problem of ascribing legal responsibility in the case of robotics

In the following, “autonomous” is used in a broad sense, meaning nothing more than a certain space for decision-making for the machine. For a project working on different understandings of autonomy and their changes because of new human–technology interactions see http://​www.​isi.​fraunhofer.​de/​isi-de/​v/​projekte/​WAK-MTI.​php. Of course, notions such as “decision” or “learning”, which I will have to use at some points in this paper, are not meant to imply that these processes are similar to human processes of deciding or learning—instead, these are always insufficient analogies (see to the criticism about the similarity argument below). Still, until now, there are no more adequate notions to describe the processes and the analogies do represent the minimum resemblance between humans and machines which is, inter alia, problematic.

Although the prognosis of Gates (2007) has not come into reality yet and might not for some time, it still cannot be doubted that service robots already have and will be of more and more relevance in households.

“Fortunately, these potential failings of man [passion for inflicting harm, cruel thirst for vengeance, unpacific and relentless spirit, fever of revolt, lust of power, etc.] need not be replicated in autonomous battlefield robots” (Arkin 2008, p. 2).

The following aspects are discussed in more detail in euRobotics (2012).

For the purposes of the Product Liability Directive 85/374, “damage” means (Article 9) damage caused by death or by personal injuries; damage to an item of property intended for private use or consumption other than the defective product itself.

Article 6 Product Liability Directive 85/374 states that a product is defective when “it does not provide the safety which a person is entitled to expect, taking all circumstances into account, including: (a) the presentation of the product; (b) the use to which it could reasonably be expected that the product would be put; (c) the time when the product was put into circulation”.

The producer is freed from all liability if he proves (Article 7): “(a) that he did not put the product into circulation; or (b) that, having regard to the circumstances, it is probable that the defect which caused the damage did not exist at the time when the product was put into circulation by him or that this defect came into being afterwards; or (c) that the product was neither manufactured by him for sale or any form of distribution for economic purpose nor manufactured or distributed by him in the course of his business; or (d) that the defect is due to compliance of the product with mandatory regulations issued by the public authorities; or (e) that the state of scientific and technical knowledge at the time when he put the product into circulation was not such as to enable the existence of the defect to be discovered; or (f) in the case of a manufacturer of a component, that the defect is attributable to the design of the product in which the component has been fitted or to the instructions given by the manufacturer of the product”.

First of all, again, the notions of adaptation and learning are meant only as analogies to human processes. Second, it is important to note that, although in many ways the problem of “many hands” (Jonas 1982) is relevant here, the possibility of learning and adaption is, in fact, the crucial difference. Because this development is inbuilt into the machine, one does not only deal with side effects of cooperation but with intended effects of uncontrolled development of machines.

See Amtsgericht München, Urteil vom 19.7.2007—Az.: 275 C 15658/07, NJW RR 2008, 40.

This is not a typical legal question; thus, it is not discussed in depth here—the background aspects would also probably require a different paper. It is mainly for politics (and ethics, political and social sciences) to discuss this question in more detail. Still, I think it should be mentioned because until now, the main focus of the debate lies on the question of responsibility after damaging a third party. For this to happen, one has to decide beforehand if robots should actually be used in a certain area of life. This question is rarely discussed—in public as well as in the academic debate—although it should, in my eyes, be the first step before discussing responsibility for damages. For further inspiration see, inter alia, Brownsword 2008.

A computer program that can take over a task for its user, having a certain amount of artificial intelligence to perform this task partly autonomously and the ability to interact with its environment rationally.

This is the case even if the user is aware of the machine’s programming and all relevant circumstances (John 2007, p. 101).

This is related to German Law: § 179 Par. 3 S. 2 BGB.

Of course, it is thinkable to even go further and develop completely new categories that do not resemble any legal concepts. For this, one would have to inquire more deeply in other disciplines and construct legal regulations from scratch—this, of course, is always a possibility. The argument here is based on the idea, though, that in legal debates one generally refers to similar situations, draws on analogies and develops the existing system further instead of creating something completely new; thus, the legal argumentation would probably at least start from existing situations and constructs one can refer to.

One possible solution is to include all unattended systems that perform tasks specific to their special character that cannot be foreseen and controlled in advance (Hanisch 2010, p. 208).