Believing in BERT: Using expressive communication to enhance trust and counteract operational error in physical Human-Robot Interaction

📝 Abstract

Strategies are necessary to mitigate the impact of unexpected behavior in collaborative robotics, and research to develop solutions is lacking. Our aim here was to explore the benefits of an affective interaction, as opposed to a more efficient, less error prone but non-communicative one. The experiment took the form of an omelet-making task, with a wide range of participants interacting directly with BERT2, a humanoid robot assistant. Having significant implications for design, results suggest that efficiency is not the most important aspect of performance for users; a personable, expressive robot was found to be preferable over a more efficient one, despite a considerable trade off in time taken to perform the task. Our findings also suggest that a robot exhibiting human-like characteristics may make users reluctant to ‘hurt its feelings’; they may even lie in order to avoid this.

💡 Analysis

Strategies are necessary to mitigate the impact of unexpected behavior in collaborative robotics, and research to develop solutions is lacking. Our aim here was to explore the benefits of an affective interaction, as opposed to a more efficient, less error prone but non-communicative one. The experiment took the form of an omelet-making task, with a wide range of participants interacting directly with BERT2, a humanoid robot assistant. Having significant implications for design, results suggest that efficiency is not the most important aspect of performance for users; a personable, expressive robot was found to be preferable over a more efficient one, despite a considerable trade off in time taken to perform the task. Our findings also suggest that a robot exhibiting human-like characteristics may make users reluctant to ‘hurt its feelings’; they may even lie in order to avoid this.

📄 Content

Abstract— Strategies are necessary to mitigate the impact of unexpected behavior in collaborative robotics, and research to develop solutions is lacking. Our aim here was to explore the benefits of an affective interaction, as opposed to a more efficient, less error prone but non-communicative one. The experiment took the form of an omelet-making task, with a wide range of participants interacting directly with BERT2, a humanoid robot assistant. Having significant implications for design, results suggest that efficiency is not the most important aspect of performance for users; a personable, expressive robot was found to be preferable over a more efficient one, despite a considerable trade off in time taken to perform the task. Our findings also suggest that a robot exhibiting human-like characteristics may make users reluctant to ‘hurt its feelings’; they may even lie in order to avoid this.
I. INTRODUCTION The new arena of collaborative robotics is sorely in need of strategies to deal with the challenges that arise where robots and humans work in close proximity. The human- populated world contains an infinite number of unknowns. Thus, while collaborative robotics is still in its nascent phase, designing for the initial mistakes, misunderstandings and failures likely to arise between human and robot is crucial. How does a robot recover a user’s trust after an error? How effective is an attempt to rectify the situation, or an apology, in mitigating dissatisfaction caused by unpredictable behavior? To what extent can a machine displaying human- like attributes soften displeasure? These are the central questions this paper seeks to begin to address. Typically, evaluation has tended to emphasize the efficiency of collaboration under varying conditions, as outlined in [27]. This (simulated) study suggests that efficiency is not the most important aspect of performance for users and seeks to demonstrate that embodied emotional expressiveness improves the integration of human-robot activity. Taking this one step further, we explored whether expressiveness - verbal and non-verbal - could mitigate any

• This work was supported in part by the EPSRC grants EP/K006320/1 and EP/K006223/1, as part of the project “Trustworthy Robotic Assistants.”
  a A. Hamacher is with the UCL Interaction Centre, University College London, UK. She is also a freelance journalist (+44 7808 719739; email: adriana.hamacher.11@ucl.ac.uk, adahamacher@gmail.com).
  b N. Bianchi-Berthouze is with the UCL Interaction Centre, University College London, UK (email: n.berthouze@ucl.ac.uk).
  c A.G. Pipe is with the Bristol Robotics Laboratory, Bristol, UK (email: Tony.Pipe@brl.ac.uk). d K. Eder is with the Department of Computer Science, University of Bristol, UK (email: kerstin.eder@bristol.ac.uk) and leads the Verification and Validation for Safety in Robots research theme at the Bristol Robotics Laboratory, Bristol, UK.
  dissatisfaction caused by erroneous or unexpected behavior and positively affect the participants’ experience.

Figure 1: The BERT2 platform with neutral expression (left) and BERT C’s facial expression on egg drop (right). Our study took the form of a real-life task, with participants spanning a range of ages and levels of experience (with robots), working directly with a humanoid robot assistant (Fig. 1) in an omelet-making task. A total of 15 of the 21 participants preferred the communicative, personable robot over a more efficient, less error prone, but non- communicative one. Satisfaction was significantly increased in the communicative condition and participants were particularly responsive to this robot’s apparent awareness of its error and expression of regret. For the majority, personable, transparent behavior appeared to negate the fact that the interaction took 50 per cent longer than in the non- communicative conditions. Our results suggest that users are likely to prefer an expressive and personable robot, even if it is less efficient and more error prone, than a non-communicative one. Our study furthermore offers fresh insight into how mistakes made by a robot affect its trustworthiness and acceptance in human-robot collaboration. It suggests that a robot effectively demonstrating apparent emotions, such as regret and enthusiasm, and awareness of its error, influences the user experience in such a way that dissatisfaction with its erroneous behavior is significantly tempered, if not forgiven, with a corresponding effect on trust. In fact, human-like characteristics may make users reluctant to hurt a robot’s ‘feelings’ and they may even lie in order to avoid this. II. LITERATURE REVIEW As robots are increasingly developed for use in social settings, acceptance, persuasiveness and likability are key and Believing in BERT:
Using expressive communication to enhance trust and counteract operation

This content is AI-processed based on ArXiv data.