Exploring task and social engagement in companion social robots: a comparative analysis of feedback types

Abstract

In recent years, the integration of social robots into various domains has received significant attention due to their potential to engage users in meaningful ways, offering companionship, support, and assistance in tasks, particularly in healthcare. This study investigates the impact of different types of feedback provided by the social robot Furhat on user engagement during a digital visuospatial memory training task. Using a $ 3 \times 2 \times 2 $ 3x2x2 mixed design (N = 58), we investigated three types of feedback: performance-based, affective-based, and a combination of both, across two levels of challenge (Easy and Medium) between subjects, incorporating a within-subject baseline control block. The results indicate that affective-based feedback leads to significantly higher social engagement, as evidenced by higher eye contact with the robot. However, this higher social engagement is associated with lower task performance in the affective-based feedback condition. Additionally, participants perceived the social robot as more user-friendly in the combined feedback condition and as more distracting within the Medium challenge level. This research provides insights into the ways in which social robots can be used to facilitate human performance and engagement in tasks where both positive attitudes towards the task and high performance are essential for long-term involvement.