Empathy-Inspired Voice Companion Design for Older Adults

Abstract

The rapid growth of the global ageing population highlights the urgent need for effective strategies to address key challenges, such as loneliness—a common issue with serious health consequences. Interventions that provide companionship by promoting emotional support, social connection, and active engagement have been shown to reduce loneliness, improve quality of life, and decrease health risks linked to social isolation. However, consistent access to human companionship is often limited by practical constraints, creating a need for sustainable alternative. In this context, recent advances in AI-driven voice assistants (VAs) have gained attention for their potential to support the well-being of older adults. Studies suggest that, over time, older adults may perceive VAs as companions, helping to ease feelings of loneliness and isolation. Despite this potential, current VAs are limited in their ability to function as true digital voice companions (VCs), mainly due to a lack of emotional depth and empathetic interaction. Research emphasises that a high level of expressed empathy is essential for older adults to perceive a VA as a genuine social companion. Yet, most existing systems lack sufficient empathy modelling, limiting their ability to build emotional connections. This paper contributes to ongoing efforts to enhance empathy in AI-based VCs. It offers a detailed review of key concepts, such as companionship, empathy, and the current state of empathetic VA technologies. Building on this analysis, the paper proposes a novel VC system design that integrates advanced empathy modelling to improve the emotional and social experiences of older adults through AI-enabled conversations.