Seite 4: Problem Areas of Service Robots

Service robots mingling among us, living together with us, sharing our routes, zones and places with us, and being in our buildings, represent risks and opportunities, and some of them relate directly to us, our creature comforts, our physical integrity, our survival, all of which already addresses moral and social aspects. Hereinafter I will focus on problems and present solution approaches, but will also point out apparently unmanageable or immovable limits.

Collisions and Crashes

Most of the mentioned service robots vary in size between animals and humans. The parcel robot by Starship Technologies is the size of a smaller dog. When moving on the pavement of cities, which is the plan of most operators, it will inevitably become an obstacle. Pedestrians, often also skaters, scooters and cyclists are on their way on the pavement. If a robot of this kind comes around the corner or is simply overlooked, serious falls may result. Thus, the physical integrity of people is at risk. Also, the robots themselves may be harmed, causing an economic damage. In both cases, there may be a question of liability.

In the case of larger robots such as the K5 or the AnBot collisions may occur. Actually, a minor accident happened at the Stanford Shopping Center in 2016, in which a boy suffered some bruises [3]. K5 and AnBot are not only relatively large, but also quite heavy, which can be detrimental for people involved in a collision. In addition, they move quite speedily so that, should the occasion arise, two moving bodies clash. In addition to the surfaces of the robot, exposed parts such as buttons and tools can cause injury.

The potential solutions lie at different levels. First, one can orient oneself on the construction of motorized vehicles. They can usually draw attention to themselves through light and sound signals by using flasher and horn. One would have to automate the activation on the models mentioned. On pavements and in public places, the problem occurs that these signals would have to be used very often, which may result in a certain tolerance on the one hand, but on the other hand may represent a nuisance, especially if you face and hear several robots at the same time. It also could scare domestic and wild animals, which subsequently might avoid these areas or run away in affect and cause damage. With security and surveillance robots, a signal is not always the first choice, because the previously informed headquarters would probably like to watch an operation without drawing attention to the robot.

One might further force the robot to move very slowly with a cruise control of some kind or a brake system, which however neither eliminates it as a stumbling block nor sorts out the phenomenon of running and driving people. In addition, the interaction of sensors - cameras, radar, lidar, ultrasonic and infrared systems, etc. - and algorithms is to be designed in such a way as to make a stop or dodging occur as early as possible as well as target-oriented. Particularly while dodging, however, the danger of new collisions arises. Last but not least, there is the possibility of modelling the robot on approaches from social robotics by creating a soft shape around its shell or make it withdraw dangerous instruments and parts if necessary.

In addition to such technical and design approaches that come from the machine itself or are connected with it, stemming from its behavior and nature, some personal, organizational and space planning measures are needed. The robot can be remote-controlled in parts - as is intended for the parcel robot of some companies, for crossing streets [14] - and one can navigate it through less bustling areas, which, in turn, can be made dependent on the time of day and information about events.

Sharing of the Habitat

As made clear already, the service robots share with us our habitat. Today, many cities are already very complex, and pedestrians, bicyclists, car drivers and animals must constantly heed several moving and stationary objects and assess them. The robots increase this complexity. We have to dodge them or they have to dodge us - as already mentioned -, they fill up the roads and squares, and make passages narrower. In other words: Our habitat dwindles further, will be further restricted, filled with artifacts that we create, and the more there are of these, the tighter it will be for us all.

Of importance are not only the places, but also the resources. Robots need electricity, often produced by nuclear and coal-fired power plants. They need this power directly, we indirectly, and since we all need this power at the same time, we are in constant competition with each other. Alternative and independent energy supplies are rare. The Energetically Autonomous Tactical Robot (EATR), a prototypical war robot, "feeds" on organic materials, which caused a controversy [8]. Apart from that, robots are produced from certain materials and substances that consume resources, though this plays rather an ecological and economical role, and hardly a role for their specific use.

It is difficult to compensate for the increase in number, space and energy demand of robots. One approach would be to make them smaller and lighter. In many cases, however, and for various reasons, service robots must have a certain extension and weight, or a certain stability. This is, on the one hand, due to their tasks like transporting food and parcels, on the other hand, because as artifacts themselves they move through artifacts (such as roads and buildings), which, in turn, have people (and possibly their pets and farm animals) as a benchmark. In this respect, neither a small nor a giant robot makes sense.

Another approach would be to roof pathways, and to create new areas in the cities to somewhat separate such machines and people, where it makes sense. Already, streets are covered with other streets, but this is very expensive and resource-intensive. Such double engagements in the double sense involve a new complexity, and, in addition to the extension, a restriction of opportunities as well.

A solution in relation to the resources would be to equip the robots (or the stations where they recharge) with a self-sufficient, sustainable, eco-friendly power supply such as solar cells. A robot with integrated power supply also has the advantage that it practically never fails, which is of the utmost importance for surveillance robots as well as for parcel robots. If they are to be effective, surveillance robots must be set in use in several versions; this is not necessary in parcel robots, except when they work together like in a relay race.

Communication and Interaction

Service robots, which move in the modern cities of the information societies must communicate and interact in multifaceted ways. They meet people of very different cultures, whose facial expression and gestures are varied, and who speak numerous languages. In addition, they meet birds, dogs and cats, which they not only have to dodge or which they scare, but with which they must also communicate and interact.

This subject has been partially addressed already in the first section of this chapter: Signals and sounds are used also for communication and interaction. In addition, natural language skills may be popular in software robots, and in hardware robots not uncommon. A transport robot could ask whether it has arrived at the right person, and ask for a linguistic or other verification. Information robots should provide structured and validated information. It is important that this information can be quickly and properly understood, and used.

Disciplines such as human-computer interaction, human-machine interaction and human-machine communication, as well as animal-machine interaction provide solutions to this set of issues. You can create the machines as adaptive systems that adjust themselves automatically to different people and languages. When a user addresses a robot in a particular language, it can answer in the same language, or enquire whether other languages are desired. It is also possible to identify the age of the person by face recognition and sizing, and communicate accordingly. The robot could also serve as an interpreter in common and public spaces, which would make it a bridge between humans and, in turn, would facilitate their understanding of each other.

Furthermore, the robot can also mimic animal noises, and by the use of signals, sounds, tools and blowing and wagging facilities attract or frighten away animals. The cooperation of machine ethics, animal ethics and animal welfare can produce systems that qualify and quantify the animals and are able to identify animal species [7], and ultimately help to improve animal welfare and reduce animal suffering.

Not to forget the direct powers of intervention, such as pressing an emergency button (which in the AnBot is already in use) and pressing a "kill switch", as suggested by a work group of the European Parliament [15]. The former allows the summons of human help, if there is danger ahead, the other to switch off the robot. Both are technical solutions, whereby the emergency button is related with organizational and personal measures. They both raise questions. What if the emergency button is being abused? What if a person without the necessary competence, who has neither the ability nor the permission to do so, presses the "kill switch"? Or, what if the machine would make the right decision and is then being thwarted?

Support and Replacement of People

A more fundamental phenomenon is that robots support people as well as replace them. Automatic and autonomous machines, as the prefix points out, do something themselves, independently, or of their own accord, and they do not necessarily need humans. These are, however, often required, when it comes to the input or output.

In the areas that are in the focus of our interest, a support or replacement has all sorts of implications. A parcel robot substitutes a postal worker when transporting; but does it substitute him or her also at the delivery, where so far have occurred social interactions, talks or a flirt? A surveillance robot supports the monitoring staff and replaces it under certain circumstances, but what happens if a suspect is to be stopped? Must the robot cooperate then peer to peer, with the monitoring personnel, with randomly present people?

These considerations lead to three main findings. Firstly, robots among us accept work from us. Thus, since these tasks are lost to us we are looking for new tasks, have to determine new ones, etc. Secondly, communication and interaction partially cease to exist or vary greatly. This topic was addressed in the previous section. Thirdly, new forms of cooperation are necessary. Partly, this subject as well has already been dealt with in the last section, since cooperation often goes hand in hand with communication.

In turn, the solutions relate to several levels. We can discuss this question in information, technology and business ethics, namely how far we will push the automation and autonomy, which tasks we intend to outsource and which we want to retain. This requires a definition of the properties and activities of machines and humans, and also a new definition of work practice and the division of labor. For the cooperation and communication with machines, new skills of machines and humans are required. The machines need to understand humans better, and vice versa.

In turn, approaches from machine ethics and social robotics can be pursued. Ultimately, machines and humans need to meet in their moral and social environment. They should behave towards each other in such ways that enable a good life and safe operation.

Data Collection and Data Analysis

Another problem has to do with data and information. Some of the robots are trimmed to recognize certain conspicuous issues, others have to find their way around in standard situations. In both cases, patterns are analyzed and transmittable images generated. This is important to improve the machines and increase their operating security. Some machines need to master emotion analysis and facial recognition. Overall, this creates a problem of data protection and informational autonomy in relation to the recorded humans, whose personal characteristics or behaviors are possibly verified and stored.

The special perspective of some service robots has also to be included in the considerations. Parcel robots may look under our skirts and between our legs [3]. They notice broken shoes, calves lined with varicose veins, and thighs rutted with cellulite. This is not tragic, but perhaps uncomfortable and embarrassing for us. Surveillance robots can be equipped with lowly located sensors, which creates a similar problem.

The solution approaches come among others from machine ethics, informatics, and AI. You can teach parcel robots not to record certain areas of the body or faces, and not to evaluate them. Such approaches were discussed in relation with photo drones [5]. Concerning surveillance robots, this is much more delicate, because they should in fact provide personal data - of a burglar or handbag robber - in case of need. But, here as well, the data can be encrypted to be decrypted only by an authorized person, or raw and application data can be distinguished, and areas of the body and body parts can be excluded.

It is important to take the perspective of information ethics and to think and act according to its terms. The informational autonomy, to name a central concept, is the possibility to access information by oneself and to look at one’s own personal data and, if necessary, make adaptions [4]. Social and political groups and entities need to refer repeatedly to this moral dimension, which is beyond the legal one. The informational self-defense - also a term of information ethics - arises from digital disobedience or represents an independent act in the heat of the moment, and is intended to safeguard the informal autonomy and the digital identity [4]. It needs to be discussed when to strike back against robots and how to protect oneself from them.