Not so long ago, it was still a vision of the future, today we are already in the middle of it: cyber-physical systems (CPS) combine mechanical and electronic components with software, allowing machines to communicate with each other. There are computer programmes, so-called software agents, which independently carry out actions within defined processes without human intervention.
Examples that nowadays already function in this way are high-frequency trading on the stock exchange or the trading of financial products by means of algorithms. Legally binding transactions can be concluded within milliseconds. This accelerates and facilitates processes enormously, but also raises certain questions.
In the next few years, especially in the area of smart production or smart factory (see box), such technological innovations will open up new opportunities for companies and help them to survive in the competitive market.
Machines can already be programmed to signal to each other, for example, what the room temperature should be so that ideal conditions always prevail for machines or server systems and to also regulate them by themselves. However, the possibilities do not end there. It will be possible for entire production processes to be carried out completely autonomously by machines without human support – like a car from the first screw to the installed software – or for machines to independently check the stock and, as soon as it reaches a critical point, automatically reorder the required components.
This scenario of a fully autonomous production environment controlled by machines is not a futuristic vision. The basic technologies are available. Machines are already able to interact with each other and in some cases no longer require human control.
The smart factory can be described as a subcategory of the Internet of Things. This umbrella term also includes, for example, smart grids – intelligent electricity networks in which generation, consumption and storage can be dynamically controlled – or smart buildings, which automatically control all technical components in a building in order to increase building efficiency. Basically, smart factory refers to a production environment that organises itself.
The economic added value of implementing these novel technologies is, on the one hand, that financial expenditure can be reduced enormously. Just-in-time orders, for example, can minimise inventory and thus costs.
On the other hand, it enables companies to do something that was previously almost inconceivable for economic reasons: to produce one-off items in a profitable way. The technology thus paves the way for industry from abstract mass production to individual one-off production.
Moreover, as already mentioned, modern sensor technologies and the networking of cyber-physical systems (CPS) also make it possible to determine and set optimal environmental and production conditions in the factory of tomorrow.
Therefore, the questions for manufacturing companies will be, first, whether and in what they should invest in this area in order to remain economically competitive. Second, there is the question of what needs to be taken into account when implementing these new technologies. An important aspect of this is the legal component.
First and foremost, two areas of law are strongly affected: Civil law, due to the possible conclusion of machine-to-machine contracts, and data protection, since the Internet of Things is characterised by the increased networking of sensors that can be installed at the workplace. As a result, more and more data from employees is being collected, linked and processed.
Basically, technical standardisation and the associated adaptation of the legal situation would be advisable. In addition, however, a harmonisation of legal regulations for operational processes is also necessary. One sticking point is the liability of companies, which can arise from claims under civil or criminal law. One thinks of strict liability under product liability law or injuries to legal assets without human intervention. The question of liability in contracts between machines should also be raised here.
In this context, it appears problematic that in legal transactions it is absolutely necessary to attribute declarations of intent (see box) to a human being. Thus, the question arises as to when the declaration of intent has been received and to whom the declaration of intent can be attributed in the case of a conclusion of a contract among machines. Moreover, the effects this problem causes in the case of breaches of contract need to be considered.
In order for a contract to be validly concluded, the declaration of intent of the contracting parties is required. According to the general rules of civil law, the declarations of intent must be attributable to a natural or legal person. Should software agents now conclude contracts for their principals, civil law must also regulate this form of representation. Here, the clarification of whether it is a matter of a messenger or a representative is of essential importance.
The two legal forms differ in that the messenger merely delivers a declaration, whereas the representative makes his or her own declaration of intent. A human act of declaration and an act of will are required for the conclusion of a contract. As a result, software agents can be regarded as messengers, since they cannot make their own declarations of intent. However, if software agents act independently within certain parameters and subsequently no direct control by a human individual is necessary, it must be assessed in each case with regard to the concrete capabilities of the system whether software agents can be classified as representatives under certain circumstances and whether, with artificial intelligence, they can even commit themselves under civil law.
Declarations of intent only become effective upon their receipt in the sphere of influence of the recipient of the declaration. Access is assumed when it can be expected that the recipient will take note of the declaration. For example, in the case of a business e-mail sent on Sunday evening, it is assumed that it will be received on Monday morning in the normal course of business from Monday to Friday. Whereas human intervention was previously required to conclude contracts, the declaration of intent now takes place within microseconds at the software agent of the recipient of the declaration. It must be expected that the declaration of intent will be taken note of immediately under normal circumstances. This leads to further problems with revocation and the contestation of error, which, however, will have to be countered less with technology and more with a further development of the civil law system.
With the new technologies for production simplification, there is also the danger and the possibility of uninterrupted monitoring of employees. Machines can record and evaluate the work processes of individual employees. Not only are work steps recorded on the factory floor by smart gloves, for example, but also at the workplace of a normal office worker by Microsoft Office 365. In this context, the Internet of Things, as already mentioned above, is characterised by a growing network of sensors that can be installed at the workplace. This also increases the probability of establishing a personal reference, even if the individual data is not personal. This can lead to data protection challenges.
The principle of transparency requires that personal data must be "processed lawfully, fairly and in a manner comprehensible to the data subject". In many jurisdictions, the Labour Constitution Act stipulates that control measures and technical systems for controlling employees that "affect human dignity" should only be permitted with the consent of the works council and be the subject of an enforceable works agreement. Thus, the processing and linking of personal data in background systems should at least be made transparent. Alternatively, or where there is no works council, in many European countries the consent of the employees is necessary, which must be given freely and without any coercion.
This article is only an excerpt from the challenges and requirements that companies and legislators will face in the future. The adaptation of the legal framework in Europe to automation within digitalisation will be of great importance in the coming years. The extent of the semantic capabilities of the machines will essentially determine the requirements for the law.