How do you achieve natural ‘social’ interactions in AI applications, for example, when robots are employed in healthcare, education and the public transport sector? This is exactly what Professor Tibor Bosse will be focusing on in his new chair at Radboud University. On 18 January, Bosse will hold his inaugural lecture, in which he will argue that this type of interaction should be facilitated through a new multi-disciplinary field of study: social artificial intelligence. We asked him three questions about his inaugural lecture.
Social robots in the healthcare sector, digital assistants and chatbots: these are just a few examples of actual developments within artificial intelligence that are having a major impact on our society. When it comes to promoting our well-being, a smooth interaction with these robots plays a significant role. But how can we ensure that this is achieved?
How do you create social artificial intelligence?
“It is difficult to predict the effect that new technology will have on people. That is why in the field of artificial intelligence we need to collaborate with fields such as psychology and social science. My area of interest forms the junction between these very fields. What exactly is it that gives people the feeling that they are talking to a human being instead of a computer? And what would it take to give people this feeling? This is not something than can be solved by the exact sciences alone, because it also requires research on human behaviour”.
And conversely: can artificial intelligence also be used to help improve our social skills?
“Yes, a number of projects are currently being carried out in this area. For example, we are working together with the RET (the main public transport operator in Rotterdam) to see how we can use a virtual reality game to teach bus drivers to deal more effectively with aggressive passengers. From January onwards, this game will actually be used in the RET training programme. I am also collaborating with KBO [The Catholic Association of Senior Citizens] to develop an app that will teach senior citizens how to recognize scams. It may happen that someone comes to the door and while they claim that they have come to read the metre, they will then raid the victim’s house. Senior citizens can use this app to enact certain scenarios. The computer can also analyse the tone of the voice and determine whether something is said assertively or not. This Dutch app can be found in both the iTunes Store and the Google Play Store by searching for ‘trucs tegen babbeltrucs’ (‘tricks to beat scammers’).”
What is required in order to realize the field of social artificial intelligence?
“We are still a long way off from seeing completely normal conversations take place between people and robots. For example, an important requirement for this is that a computer must be able to explain how it came to a particular conclusion. And in order to achieve this, greater investment needs to be made in two areas: AI techniques, such as machine learning, should be further developed and more attention should certainly be paid to multidisciplinary challenges, which include social artificial intelligence”.
A Special Interest Group, which was founded by BNVKI (Benelux Association for Artificial Intelligence), has recently launched a manifesto that describes the direction in which Dutch AI research should be heading. Tibor Bosse is chair of BNVKI and a joint initiator of the manifesto, which can be found here(verwijst naar een andere website).
Tibor Bosse will hold his inaugural lecture on Friday, 18 January in the Radboud University Auditorium. Please visit these links for more information about his inaugural lecture(verwijst naar een andere website) and registration(verwijst naar een andere website).
Please note: the registration page is only available in Dutch.
Would you like to know more? Please contact:
Tibor Bosse, +31 (0)24 361 2018, t.bosse@ru.nl
Science Communication Radboud University, +31 (0)24 361 60 00, media@ru.nl
Radboud AI is a platform of Radboud researchers with various scientific backgrounds, from linguists to data experts, who work together to develop ‘responsible AI’: next generation AI. In this process, Radboud AI chooses the human dimension as a common thread. For more information, go to www.ru.nl/ai
Comments