Machine learning provides huge amounts of data with learning algorithms and machines then make autonomous decisions. The best example of this is self-driving cars. But those who programme the learning algorithms or determine the training methods for the AI also always influence how the machines process these algorithms. For example, one family man will be more likely to prevent the self-driving car from hitting children in an accident. Another may decide to save the pensioner. Thus, it can be proven beyond doubt that "technical solutions always incorporate the values of those who invented them and those who finance them" (quote Dr. Thorsten Busch).
Ethics as the "doctrine of the good" is obliged to point out principles and values with which to reflect on concrete, area-specific questions. Finally, it is also a question of which intellectual instruments and methods we want to use to meet the new challenges. Through the digitalisation process and the development of Eliona that we have gone through in recent years, new options for action have emerged in relation to
Mass data collection > IoT sensor technology
The essential principle of the networked world is to capture more and more facts digitally. However, this simultaneously raises the question of the limits of what can be digitised.
Data evaluation > Big Data
With Big Data, we look for correlations and their evaluations. However, the functioning of the evaluation algorithms usually remains in the dark. This raises questions about the significance, manipulability and revisability of certain algorithm-based decisions.
Interpretation of the data > Artificial intelligence
Increasingly, decision-making situations in which humans traditionally acted are now being taken over by machines.
So who bears the responsibility in our case for a failed interpretation of the data by Eliona?
Many ethical questions that seem new to us in the guise of digitalisation are in fact not new. For example, the now very famous dilemma of whether a networked car should run over a pensioner or a child in an emergency is by no means a new phenomenon that is only being discussed in connection with automated driving. Basically, it is a question that is more than 2,500 years old and that has been pondering mankind for generations under the name of "Board of Carneades", among others.
Algorithm-based Big Data decisions, for their part, are strongly reminiscent of the well-known phrase: "Don't trust any statistics you haven't falsified yourself." And we have also been familiar with questions about the relationship between man and machine since the beginning of industrialisation. In discourses on net ethics, it is therefore often worthwhile to look first at the analogies that have long been discussed as ethical problems. For it is through the potential of the Internet of Things, manifested through platforms like Eliona, that we are able to create advances and facilitations.
This makes it possible, for example, to manage infrastructures more efficiently, control energy flows and provide resource-saving air conditioning for workplaces. The future is moving inexorably in this direction and it is important to address ethical principles at an early stage. Especially because we ensure the safety of people in buildings and on transport routes through the transmission of information. The subject area includes operational safety (Safety) and personal safety (Security) in both the real and the virtual world.
With platforms like Eliona, an increasingly ageing population will be able to live longer in their own homes. New and liveable forms of cohabitation between generations are emerging and a variety of services are being made possible. User groups are not only the older generations, but also single parents, singles and normal families with children. Eliona has such a modular structure that it is possible to define exactly what should be networked and what data we want to transfer from our daily lives. Data protection plays a decisive role in the platform. You have to define in detail what should be collected, what should be evaluated and to what extent the AI is allowed to "machine learn" this data.
The human image should be the yardstick for ethical reflection. Ethical norms and standards for "good" behaviour can be derived from an understanding of what actually constitutes a human being.
Author: Björn Erb