The browser you are using is not supported by this website. All versions of Internet Explorer are no longer supported, either by us or Microsoft (read more here: https://www.microsoft.com/en-us/microsoft-365/windows/end-of-ie-support).

Please use a modern browser to fully experience our website, such as the newest versions of Edge, Chrome, Firefox or Safari etc.

Robots – not so smart as we would like to think

Man with robot
Christian Balkenius is not worried about robots taking over and becoming smarter than us. " Not in my lifetime, in any case", he reassures. Photo:Maria Lindh

How do you get a robot to behave in an ethical and moral way?
Christian Balkenius is giving this a lot of thought, as it is the topic of his research project. However, he is also thinking about ethics among robot researchers. 
“It’s often said that we have advanced further than we actually have done. The aim of the research is not fulfilled if we are not honest.” 

Robots pique the imagination and society’s expectations have been heightened by the media. It is not always easy for researchers to resist and sometimes reporting is exaggerated. The same thing may happen when applying for grants.

“It’s a case of really wanting to appear to have progressed a lot further than the last time you applied for funding… “

There is also a rush to publish due to the nature of the qualification system. And there is a tendency to exaggerate the significance of their research results, comments Christian Balkenius.

He is Professor of Cognitive Science and Director of the graduate school in the WASP-HS research programme (Wallenberg AI Autonomous Systems and Software Programme – Humanities and Society) which will engage 70 doctoral students throughout Sweden in due course.

He is a central figure for the biggest investment ever in the humanities and social sciences in Sweden. At present, the funding of SEK 660 million covers 16 research projects (three of which are in Lund) and the graduate school. The unifying theme is how artificial intelligence is going to affect people. The Wallenberg Foundations are providing the funding and that has a special significance according to Christian Balkenius. 

“They have funded a lot of the AI research in Sweden and therefore it’s pleasing that that they now want to fund something that is about the consequences”, he says. 

So, where does robot research stand – will the robots take over and become smarter than us?

“No, no, not in my lifetime in any case. We are only at the beginning of this development”, reassures Christian Balkenius.

That does not mean, however, that technological development is standing still. On the contrary, there is a lot happening and relatively fast. However, there is a long way to go before a robot can reach a human level. 

“Nowadays, a robot can function in a factory or in a spacious warehouse – or in a large office with no carpets or thresholds. They can move paper and things from one place to another.” 

Christian Balkenius has a couple of robots at home, but he is only content with the robot lawnmower. The robot vacuum cleaner is not doing a very good job… 

“A lot of intelligence is required to function in a home.”

What about robots as company for the elderly? He immediately dismisses this as a bad idea. However, he sees a strong future for robots in assisting the elderly who want to remain in their own homes, and in elderly care as a whole. 

 

A friendly face is important for a robot

The smiling robot face that greets visitors to the lab at LUX, where Christian Balkenius and his colleagues work, is for research, in this case basic research. The researcher’s everyday work here is devoted to his own project “Ethical aspects of autonomous AI systems”, which he hopes will pave the way for empathic robots.

“We have started by giving them a friendly face!”

That seems a lot easier than giving them a good heart. Christian Balkenius talks about the jungle of rules that exist for how people are to act and behave in various situations.

“But robots can interpret rules in different ways and it’s not at all certain that the results will be the ones you intended”, he says.

Morality is not an altogether easy aspect either. He points to the example of self-driving cars. 

“If a crisis situation arises in traffic, the car must decide who to save – the two in the street or the person driving the car?”

Christian Balkenius tests various ethical theories on the robots and observes how they work in practice. One theory may concern duty – doing the right thing – and another on doing good. There can be a lot of ethical dilemmas. 

Research involves helping robots to try to understand how people are and how they feel

The development of artificial intelligence has come so far that it is impossible to slow it down and Christian Balkenius therefore considers that the best possible outcomes must be pursued in this situation. Asia, and Japan in particular, has established a slight lead, but now – thanks to major EU projects – Europe is in the process of catching up. There are, however, a number of cultural differences regarding views on when and how robots should be used. For example, nations such as Japan and USA are responding differently to the growing care needs of an aging population. 

“For the Japanese, solving the problem with robots is the natural path, whereas in the USA cheap labour is used instead.” 

As Christian Balkenius and his research colleagues’ work involves helping robots to try to understand how people are and how they feel, they learn a lot themselves about people’s behaviour in different situations. 

“For example, what is the acceptable distance when you stand next to another person so that it feels comfortable? And how do you adapt to different groups and sense social norms?”

It is clear that the development of empathic robots requires many disciplines, and Christian Balkenius has a background in mathematics, linguistics and computer science ­– and then a little philosophy and psychology as well.

“Most of us in cognitive science have a motley background. But, we are all interested in how thought processes work.”

 

 

 

 

 

 

GRADUATE SCHOOL IN WHASP-HS

The first of the graduate school’s 70 doctoral students around the country will be recruited in the spring and start in the autumn of 2020. Lund University will host the project. Christian Balkenius will be the director of a small organisation in Lund, which as yet only includes Professor of Theoretical Philosophy, Ingar Brinck.

The WHASP-HS Research Programme

It currently consists of 16 projects, three of which are in Lund, and each project has one to two associated doctoral students. All the projects are interdisciplinary and involve participants from the humanities, social sciences and engineering. The entire programme is organised from Umeå University.

 

 

Christian Balkenius's Project

It is called “Ethical aspects of autonomous AI systems” and includes Ylva von Gerber, Frank Zenker and Birger Johansson plus two doctoral students. The project is funded for five years. The Wallenberg Foundations are funding all the projects.

cover photo

About LUM

The first edition of Lund University Magazine – LUM – was published 1968. Today, the magazine reaches all employees and also people outside the university. The magazine is published six times per year. Editor Jan Olsson.

LUM website in Swedish

Editorial staff

Jan Olsson


046-222 94 79

jan [dot] olsson [at] kommunikation [dot] lu [dot] se

 

Minna Wallén-Widung

046-222 82 01


minna [dot] wallen-widung [at] kommunikation [dot] lu [dot] se