13:30-13:40 Introduction (Aurélie Clodic)

13:40-14:05 Psychological Foundations of Commitment

Francesca Bonalumi - Department of Cognitive Science, Central European University, Budapest, Hungary

Investigating commitment from a psychological perspective entails dealing with a major challenge: to understand which are the minimal conditions that trigger a feeling of commitment and the understanding that a potential partner is committed. We claim that by committing, an agent influences the expectations of her partner, engages her own social dispositions, and implicitly accepts to put herself in a position of being accountable for any potential defection. The empirical way to assess whether a commitment to do X is made is to measure the committed agent’s tendency to do X in spite of some material benefits for not doing X, and moral disapproval from her partner if she does not do X: these measures are made by presenting to participants several hypothetical situations in which elements such as the intentions of the committed agent, the consequences of her not doing X or the common knowledge of expectations vary. We test the hypothesis that the level of moral disapproval of an agent when a commitment is broken by a partner depends not only on the consequences of the breaking, but also and decisively on the perceived intentions of the partner, and her concrete attempts to prevent or minimize the loss that the agent is bound to suffer. Furthermore, we test the hypothesis that moral disapproval is not decisively sensitive to whether an agreement is explicit but whether there is common knowledge about the expectations of the agents.

14:05-14:30 Human-Robot Interaction: Needs and Limits 

Raja Chatila - Institut des Systemes Intelligents et de Robotique (ISIR), Sorbonne Universites, UPMC Univ Paris 06, CNRS, 75005 Paris, France.

The field of human-robot interaction is flourishing, enabled by new results in several areas such as perception, manipulation, action planning and control, inspired by theories about human-human interaction, and pulled by new needs for social or collaborative robots. The broad area of cognitive HRI addresses challenging issues such as detecting facial expressions and emotions, exhibiting emotions, understanding social signals such as postures and gestures, interpreting human activities and behaviors, detecting and reasoning on intentions, perspective taking, space sharing, cooperative and joint action planning and execution, etc. Several projects target specifically humanoid or android robots, motivated by the belief that such robots would be able to achieve more natural interactions with humans because their shape facilitates easier imitation of human gestures and expression of emotions. We shall overview a few of these cognitive capacities and consider their actual level of development, discussing how much they can actually be operational in the HRI context.


15:00-15:25 Commitment management in Human-Robot Interaction – an overview

Aurélie Clodic - LAAS-CNRS, Université de Toulouse, CNRS, Toulouse, France

In this talk, we will review some already existing works concerning commitment management in human-robot joint action. Which aspects have been taken into account and how, for example how commitment to do an action together is obtained at the beginning of an interaction. Which tools have been used to mediate this important aspect of joint action and how it is handle alongside during the action.

15:25-15:50 The Sense of Commitment in HRI

John Michael - Department of Philosophy, University of Warwick and Department of Cognitive Science, Central European University

There is a vast potential for robots to assist humans in joint actions in many different domains, from disaster relief to health care, education, and manufacturing. As roboticists move forward in optimizing human-robot interactions in order to tap this potential, it may be fruitful to consider one particular question pertaining to the challenge of designing robots with whom humans can interact comfortably and productively in various kinds of joint action. Specifically: Is it possible to design robots that elicit and/or exhibit a sense of commitment – i.e. such that (i) humans agents are motivated by a sense of commitment toward them, (ii), human agents expect them to be motivated by a sense of commitment toward human agents, (iii) they are motivated by a sense of commitment toward human agents, and/or (iv) they expect human agents to be motivated by a sense of commitment toward them. To the extent that it is feasible to implement one or more of these forms of a sense of commitment in human-robot interaction, this could be useful insofar as it could enhance human agents’ willingness to rely on robots (e.g. to perform actions which depend upon a contribution from a robot and otherwise entail a risk), could motivate human agents to perform important actions which they otherwise might not perform (e.g. to practice new skills which they are learning with a robot collaborator, to take their pills regularly because they have made a commitment to their robot nurse that they will do so, or more generally, to promote therapeutic interactions – especially with regard to elderly people, cf. Mordoch et al., 2013). In a nutshell, a sense of commitment could be highly useful in optimizing human-robot interactions, since the sense of commitment serves important functions which are of great benefit in human-human interactions. In this talk, I will first lay out some conceptual preliminaries pertaining to commitment and the sense of commitment, to the social functions thereof, and to the ways in which commitment and the sense of commitment can be generated and sustained in human-human interactions. Next, I will identify several concrete factors which may be useful for implementing the sense of commitment in human-robot interaction.

15:50-16:15  Commitment, Cooperation and Coordination in HRI

Henry POWELL - Department of Philosophy, University of Warwick, UK

Commitments are often thought to be instrumental in galvanising certain kinds of human social activity. Specifically, they are thought to play essential roles in the coordination and cooperation of individuals that makes group activity as effective as it is. It is, however, not entirely clear how exactly these three things are supposed to fit together. That is, what kinds of relations we should understand there to be between commitment, coordination, and cooperation – if any at all. Resolving this problem has obvious benefits, not only in allowing us to understand the structure of these kinds of human interactions but also by providing us the grounds from which we might be able to engineer similarly effective interactions between humans and robots. In this response I would like to raise some questions about how commitments should be thought to fit into our understanding of both coordination and cooperation and similarly, how coordination and cooperation should be thought to fit into one another. Having done this, I will consider how understanding the relation between commitment, cooperation, and coordination might benefit the development of robots designed to interact with humans in social, palliative, and industrial spheres.

16:15-16:30 Panel Discussion

Online user: 52