‘Ethics for robotics requires a tailor-made approach’

Jeroen van den Hoven, professor of Ethics and Technology, discusses intelligent systems

By Jurjen Slump

The arrival of intelligent robots on the streets, in our workplaces and in our homes has sweeping moral and ethical implications. At Delft University of Technology, Jeroen van den Hoven and his team are doing research on the question of how people can retain control of these systems. ‘A society with robots and artificial intelligence requires new ethical models.’

Imagine that your elderly mother has an intelligent device that you can use to keep an eye on her. It initially sounds like a handy aid. But as the child, you may end up feeling obliged to constantly check whether your mother is still alright. For her part, your mother also becomes aware that her child feels constantly responsible to check whether everything is still working.

Jeroen van den Hoven offers this example as an indication of the ‘subtle implications’ that robots can have on the ‘fabric of human relations and moral dependence’. ‘That means that the introduction of intelligent devices will always have moral implications.’

Of equal importance is safeguarding human dignity after the arrival of robots. ‘A secretary was recently in the news who had been replaced by a tablet. The woman was not only angry, she also felt humiliated’, Van den Hov- en says. ‘It is now becoming very real. This is now going to happen. There are an increasing number of examples of people who not only fear that they will be replaced by a robot, but actually are.’

These are somewhat philosophical questions. ‘Is this a good society? And if we want a society in which robots play a significant role, what does this society look like, and how do we create the robots?’

Meaningful human control

In this regard, a primary concern is the so-called ‘meaningful human control’ concept. Together with Aimee van Wijnsberghe and Filippo Santoni de Sio, Van den Hoven is researching this subject. ‘If robots are given increased autonomy, what does it mean to have meaningful human control of them?’ To conceive this, the concept of ‘mean- ingful human control’ – which has its roots in international law – needs to be translated to the world of robotics. ‘You see, the point is: you know absolutely nothing until you have conducted a thorough analysis of what it means in concrete cases. Because if you don’t know what it means, you cannot build anything that meets the requirements.’

Van den Hoven believes that this issue will now become palpable in robotics. ‘It is no longer sufficient to hold an ethical discussion in which terms like privacy, liability, democracy and jus- tice are bandied about without being able to say, “Okay, but that will impact how we make this or that system”. Because otherwise, you may be doing good things, but you are not actually helping the world to advance.’ In a high-tech society, with robotics and artificial intelligence, ethics will need to take a different form. It needs to be about ethics in the form of system requirements. ‘If you talk about human dignity, that needs to be translated concretely for each individual situation. For example, how to you safeguard this in a new data system for hospitals? Literally, how will that look?’

Delft Design for Values

And that is precisely what is being done at Delft: building bridges be- tween these two worlds. We already have the TU Delft Robotics Institute, and that will be joined by another institute, Delft Design for Values. This new institute is focused on building that bridge: translating existing ethical models into new applications. ‘You continue to question what it means for a certain system in concrete terms – with regard to privacy, for example. You will then eventually be left with a list of requirements that can actually help engineers during the construction process.’

Van den Hoven has high expectations of the collaboration with the TU Delft Robotics Institute. ‘We have secured a subsidy for research into autonomous mobility and meaningful human con- trol, a project involving both people from our faculty and from the Robotics Institute.’

Alongside the technicians, the opinions of behavioural scientists, social scientists and other experts are all required in order to answer the question of what is ethically the best course of action. ‘That is what we call comprehensive engineering: you have to unite the humanities, hard sciences and life sciences, otherwise you will never reach a decent solution.’