“We shouldn’t humanize robots too much”

Professor Nicole Krämer heads the Department of Social Psychology, Media and Communication at the University of Duisburg-Essen. Together with her team, she has investigated how humans treat human-like robots. In view of the increasing automation in industry, this research is relevant for cross-sector areas such as logistics.

Professor Krämer, what were the results of your research?

We are investigating how humans react to and interact with robots. For me, it is always surprising to see the results of these studies: In interaction with the machine, humans behave just like in interpersonal communication. So they are polite, speak clearly and smile when the robot smiles at them.

Prof. Dr. Nicole Krämer, University of Duisburg-Essen

“People should always be aware that robots are machines. ”

What does that mean?

Humans cannot deny that they are social beings. As soon as a robot has visible cues, such as a face, that convey human resemblance at first glance, humans react accordingly. It is astonishing, for example, that the test subjects feel flattered when they receive a compliment from robots. People are not necessarily aware of this.
It becomes interesting when it comes to switching off talking robots. People react extremely sensitively to the machine’s response. It is difficult for them to switch off someone who has spoken to them. That’s not how you behave when interacting with other people.

Today, for example, robots are used in logistics handling centers. What are the consequences of your work for those who introduce robots in their everyday work?

Artificial Intelligence launches a revolution in transport and logistics | logistik aktuell

Survey article on Artificial Intelligence, its significance for logistics and possible applications within logistics.


You have to be very careful how much you equip robots with social cues like a human-like figure or a face, because strong cues in humans can lead to confusing reactions. This is problematic because robots are not living beings and should not appear to be so. Sure, this is not a big topic today, but in a few decades it could be different. It is important from an ethical point of view that it always becomes clear that a robot is an artificial being.

How should humans react to their machine colleagues?

People should always be aware that robots are machines. In a team, the perception of a robot depends on its autonomy and capabilities. The employees in the warehouse also notice that robots are still stupid at the moment. They can’t do anything but drive fast along the shelves and lift heavy things. Nothing works without human intelligence and control.

Now everything revolves around robots that perform tasks. What about the machine that stands next to the employee in the warehouse during the break and offers a chat?

I don’t think machines need such skills at the workplace. Employers don’t want robots that take away working time.
But the topic is psychologically very interesting. There are now many considerations for such companion robots, which not only bring a glass of water, but are also available for small talk. From a psychological point of view such abilities help people without social contacts. However, we should make sure that the robot stimulates social interaction. In old people’s homes or nursing homes, this can mean that it not only signals to the elderly: XY is online right now, but then he makes an appeal: “Don’t you want to Skype with XY?

“Professor Nicole Krämer from the University of Duisburg-Essen warns that #robots must not resemble humans too much. Otherwise it will be difficult to communicate. DB #Schenker #Automation“

Tweet WhatsApp

What are you currently researching?

At the University of Duisburg-Essen, we are investigating how communication with robots changes the way people communicate with each other. It’s also interesting how much information people need about which algorithms the machines use and how they work. This is a multi-year project supported by the VW Foundation. Basically, it is about the awareness of communicating with a machine. The question is: Is it helpful for machines to explain more clearly what algorithms they work with, what capabilities they have and what happens to the personal data of human users? The latter in particular is often suppressed.

Why are you personally interested in this area?

I have been researching human-machine interaction for 20 years and am examing the non-verbal forms of communication for 20 years. I am very interested in how people communicate and react when communication is different. We learn a lot about our social nature. And of course: If we understand humans better, we can also improve robots.

Do you think that the interaction with a human-like robot will eventually no longer differ from that of a human?