380
V. Vouloutsi et al.
mation lead to a successful group selfie task?”. To answer these questions, we
performed three experiments. In doing so, we hope to further advance the field
of social robotics by contributing to the research area of social sensorimotor
contingencies and determining their effectiveness as an approach to design living
machines with social and transparent capabilities from a human-focused perspective.
1.1 Social Sensorimotor Contingencies and Personality in Social
Robotics
Social cognition can be considered as an experience generated by action, and
context is paramount (the same actions can have different meanings and consequences in different contexts). While internal reflection about the mind-state
of others occurs in several interactions, the meaning of many social interactions
is the result of dynamic relationships formed by participatory sense-making,
which allows two or more agents to generate the meaning of the social interaction together through dynamic coupling [5]. Many everyday social interactions
(like handshakes, or hugs) provide examples of participatory sense-making. This
enactive approach highlights the role of embodiment in sense-making without
intermediary construction of internal representations. Indeed, interaction detection, in its basic sense can be depicted by reciprocal sensorimotor contingencies
[8].
Sensorimotor contingencies in Human-Human Interaction (HHI) provide a
working framework from which to model and analyse basic and higher-level
behaviours (like personality) in social robotics. Personality is an important facilitator of HRI [9], and by definition, it characterises specific sets of stereotypical
behaviours that help to predict an other’s actions or make sense of their actions in
a particular social context. In HRI settings, an individual’s personality not only
affects how the robot’s personality is perceived but also the preferred personality
of the robot in social interactions. Personality preferences are modulated by both
task and group settings. In dyadic settings, extroverted robots are perceived as
more intelligent, whereas introverted robots are perceived as more intelligent in
group settings [10]. However, the evaluation of personality in social robots across
various settings and groupings is highly fragmented [9]. A widely used metric to
evaluate personality is the Big Five model [11], and extroversion/introversion is
the most commonly assessed personality trait in HRI. As personality remains a
major research area in HRI and is a known factor that affects social interaction,
we included it as a factor to this work.
1.2 The Social Significance of Selfies
Most HRI research has focused on dyadic, one-to-one interactions, and few studies have focused on one-to-many group interaction settings. As the task domain
affects the preferred and experienced personality of robots, we chose a social
setting with an inherent group dynamic that is necessary to investigate the
behaviours of the robot and their success in coordinating group interactions.
V. Vouloutsi et al.
mation lead to a successful group selfie task?”. To answer these questions, we
performed three experiments. In doing so, we hope to further advance the field
of social robotics by contributing to the research area of social sensorimotor
contingencies and determining their effectiveness as an approach to design living
machines with social and transparent capabilities from a human-focused perspective.
1.1 Social Sensorimotor Contingencies and Personality in Social
Robotics
Social cognition can be considered as an experience generated by action, and
context is paramount (the same actions can have different meanings and consequences in different contexts). While internal reflection about the mind-state
of others occurs in several interactions, the meaning of many social interactions
is the result of dynamic relationships formed by participatory sense-making,
which allows two or more agents to generate the meaning of the social interaction together through dynamic coupling [5]. Many everyday social interactions
(like handshakes, or hugs) provide examples of participatory sense-making. This
enactive approach highlights the role of embodiment in sense-making without
intermediary construction of internal representations. Indeed, interaction detection, in its basic sense can be depicted by reciprocal sensorimotor contingencies
[8].
Sensorimotor contingencies in Human-Human Interaction (HHI) provide a
working framework from which to model and analyse basic and higher-level
behaviours (like personality) in social robotics. Personality is an important facilitator of HRI [9], and by definition, it characterises specific sets of stereotypical
behaviours that help to predict an other’s actions or make sense of their actions in
a particular social context. In HRI settings, an individual’s personality not only
affects how the robot’s personality is perceived but also the preferred personality
of the robot in social interactions. Personality preferences are modulated by both
task and group settings. In dyadic settings, extroverted robots are perceived as
more intelligent, whereas introverted robots are perceived as more intelligent in
group settings [10]. However, the evaluation of personality in social robots across
various settings and groupings is highly fragmented [9]. A widely used metric to
evaluate personality is the Big Five model [11], and extroversion/introversion is
the most commonly assessed personality trait in HRI. As personality remains a
major research area in HRI and is a known factor that affects social interaction,
we included it as a factor to this work.
1.2 The Social Significance of Selfies
Most HRI research has focused on dyadic, one-to-one interactions, and few studies have focused on one-to-many group interaction settings. As the task domain
affects the preferred and experienced personality of robots, we chose a social
setting with an inherent group dynamic that is necessary to investigate the
behaviours of the robot and their success in coordinating group interactions.