Philosophy
Human identity continues to be discussed from multiple standpoints. “What is me?” has been of a central question of interest of diverse scientific fields from ethics to epistemology, from psychology to neuroscience. A supplemental question is “What makes me me for the people with whom I am interacting?”. This question takes on various different angles when addressed in the context of artificial identity. What makes it genuine, or affects the perception of it being so? On the contrary, when do we consider an artificial identity to be the product of deception? Accordingly, while artificial identity can support users’ engagement with the agent and contribute to establishing meaningful relationships with these agents, it can also mislead people, which poses serious ethical questions.
Affect
An individual’s identity is rooted in their sense of belonging, and is communicated via the expression of emotions and information. Emotions are particularly important in HRI for conveying emotions, as robots would often use emotion-related characteristics and display of emotional matters for conveying that a robotic machine is more than just an object. These displayed emotions will have substantial influence on how a robot and its identity are perceived. For example, previous studies show that robots that express positive emotions tend to be more accepted as technology than robots that do not. Speech is a central modality to convey emotions and information about one’s identity to others. Prosodic features of speech like speech rate, pitch, or loudness express information about who we are, how we feel, and our place in society. Numerous studies show that one’s speaking style influences how others behave toward them because it is associated with subconsciously encoded stereotypes. We see similar procedures when interacting with artificial agents, as humans perceive and behave differently to artificial agents that demonstrate different social cues during speech interactions.
Design
Voice design in HRI is gradually becoming more prominent but still remains at an embryonic stage. Nevertheless, considering the recent advances in speech technologies, robotics and artificial intelligence, robots’ speech is becoming remarkably closer to human speech. This is raising important questions regarding how these human characteristics influence the construction of artificial identity: how it is communicated, and in turn, perceived by humans. For example, recent reports show that the use of female voices as a default is inherently linked to an increase of verbal abuse towards the agent, and might see “real-world” consequences with verbal abuse towards women. Researchers within the HRI community have responded to these potential social consequences by experimenting with the effects of deliberately gendering robots (in both appearance and voice) on perceptions of robots and biases about people. Other work has suggested that designing agent voices to match personal preferences or even remind users of their friends’ voices might provide a sense of comfort when interacting with an unfamiliar robot or in an unfamiliar setting. Since robots are not human-beings, it is crucial to question the voices given to them, as well as how they utilize those voices, to further understand artificial identity and its implications for human-human and human-agent interactions. Hence, designing robot identity and emotion via voice and speech takes on both social and engineering challenges that should be further investigated.