Introduction

While creative and innovative product ideas traditionally originated mainly from the designers’ capabilities and creative input, more recently, the focus shifted towards the needs and goals of the end-users by following a human-centered design process and resulted in creative and innovative solutions and products. Co-creation is a method that involves consumers and stakeholders working together to create and develop new products and services (Ramaswamy & Ozcan, 2018). These collaborations are normally done in sessions, where a trained facilitator is also present to guide the group and manage the group dynamic. Additionally, the facilitator could affect the group's productivity, the total of the members' abilities, as what the group can collectively produce, such as the unique ideas resulting from a co-creation session (Frank & Anderson, 1971). However, optimal facilitation requires training, knowledge and skill, which is scarce. A social robot is one alternative to a human facilitator, as social robots bring prominent possibilities within co-creation sessions. Research has shown that social robot facilitation could positively affect productivity and creativity compared to other technologies (Alves-Oliveira et al., 2020; Kahn et al., 2016 ). Additionally, a social robot can be designed to resemble varying degrees of human likeness in terms of appearance and behaviour, as this can be done by applying anthropomorphic design features. However, applying anthropomorphic design features could result in positive and adverse effects on how people perceive the robot. 

 

It is unclear how applying an anthropomorphic design feature on a robot facilitator will affect the productivity of the co-creation session and participants' attitudes towards the robot facilitator. My master's thesis investigated how applying the anthropomorphic design feature of conversational voice (human-like) on a robot facilitator will affect their facilitation role compared to a robot facilitator with a robot-like conversational voice in a co-creation session. These effects were measured by the productivity of the co-creation sessions. In addition, the participants' trust in the robot facilitator, the participants' acceptance of the robot facilitator, and the perceived empathy of the robot facilitator were also measured and investigated whether these had mediating effects on the productivity of the co-creation sessions. 

​

Method

A lab study with a between-subjects design was conducted to address the research question: “how does the applied human-like conversational voice on a robot facilitator affect their facilitation role compared to a robot facilitator with a robot-like conversational voice in a co-creation session?”. In each co-creation session, two participants interacted with either a robot facilitator with a human-like conversational voice or a robot facilitator with a robot-like conversational voice (see Figure 1, Figure 2). The participants were randomly assigned to one of the two conditions. In total, 58 participants participated in this study. 

 

In the co-creation session, participants generated ideas regarding the problem of "What does a futuristic school in 2050 look like?" as most of the participants were students. The participants wrote down their ideas on Post-Its or an individual paper. The robot facilitator was not autonomous, resulting in the use of the Wizard-of-Oz method. However, a protocol was used that contained pre-defined sentences and guidelines on when to use those sentences. Additionally, the use of the protocol was to ensure consistency between different sessions and make participants believe that the robot facilitator was acting autonomously. After the co-creation session, the participants had to fill in a questionnaire regarding their demographic information and how they perceived the robot facilitator. The productivity was measured by counting the number of generated ideas of the participants, while trust, perceived empathy and acceptance were measured by various validated scales within the questionnaire.