Institutionen för naturvetenskap och teknik

AASS Seminar - Real AI is Social aware AI

04 mars 2019 13:00 L153, Långhuset


For more information about the AASS Seminar Series, please contact:
Alessandro Saffiotti

The research centre AASS arranges a seminar with Frank Dignum, Delft University & Umeå University.

Abstract

Although the discussion on the consequences of current AI developments are centred about the tasks and decisions that AI systems are taking over from humans a far more important issue is the way AI systems are changing our way of interacting with the world and other people. When AI is taking decisions with huge consequences, like with medical diagnoses, mortgages, autonomous driving there is often not one optimal solution, but we have to carefully determine what optimality means in these contexts. We need decisions that are acceptable, which means that the AI system needs to be socially aware and socially capable. This development will put new requirements on the way AI systems should interact with humans and even the way they are organized and the type of issues they prioritize. E.g. a medical diagnose support system, such as Watson, does not necessarily have to give an optimal diagnosis but should give an acceptable diagnosis that is correct. Such a different perspective has large repercussions on how the design and architecture of the AI systems should be approached and which are the most important issues to solve. When done properly it also leads to new opportunities for AI systems in industry and society. In this presentation I will illustrate the claims and give directions for these new research directions.

Speaker´s bio

Frank Dignum got his PhD at the VU in Amsterdam. Immediately after that he went to Swaziland (Africa) to set up the computer science department. After a year in Portugal at the Technical University of Lisbon, he spent almost eight years at the Technical University of Eindhoven before coming to Utrecht University in 2000. Since 2013 he is a honorary principal research fellow of the University of Melbourne and since 2017 also visiting full professor of the Technical University of Prague. He has been the technical coordinator of two successfully completed EU projects (MEMO and ALIVE). He is well known for his work on norms and agent communication and for socially intelligent agents in serious games. Currently these more social agents are also used in social simulations to support policy making. He has organized many workshops and conferences (amongst which AAMAS and ECAI) and given tutorials and invited presentations at most major conferences and summer schools. Since 2014 he is a EurAI fellow. Has been in the IFAAMAS board and just started the International Journal on Socio-Cognitive Systems. At the moment his H-index is 52.