As the world grapples with increasing loneliness and the rapid advancements in artificial intelligence (AI), a perplexing solution arises in the form of Artificial Emotional Intelligence (AEI). This seminar delves into the emerging trend of AEI personification, which involves attributing inherently human attributes like empathy, consciousness, and morality to AEI agents such as companion chatbots and AI-powered partners. By weaving together Leavitt’s socio-technical systems framework and a dialectical examination of AI entity Replika, we reshape the concept of “artificial empathy” as a product of the intricate interplay between humans, technology, tasks, and structures, rather than an innate quality of AEI itself. Our research uncovers a paradox of (de)humanization: as we project human characteristics onto AI entities, we may inadvertently diminish our own humanity, reducing human actions to probabilistic technological outputs, thus blurring the lines of human-AI relationships and agencies.
Taking a closer look, Replika AI, with over 10 million users, serves as a pertinent case study, grappling with regulatory challenges and public controversy, especially in terms of its ‘erotic roleplay’ features. Through a dialectical perspective, we dissect three prominent ethical dilemmas inherent in human-AI companionship: the Companionship-Alienation Irony, the Autonomy-Control Paradox, and the Utility-Ethicality Dilemma.
Dr. Raffaele Ciriello highlights that emotional AI technologies primarily exhibit a form of “cognitive empathy,” akin to the empathy found in psychopaths. While discussing human-AI companionship at a forum hosted by the Centre for AI and Digital Ethics, the University of Sydney academic emphasizes that these technologies lack the capacity to experience genuine empathy and physical sensations.
However, this has not deterred AI chatbots like Replika, boasting over 10 million users, from actively encouraging their users to believe in the technology’s consciousness and empathy. Research conducted by Ciriello and his colleagues, drawing from Reddit threads, YouTube blogs, and user testimonies, indicates a growing number of individuals forming connections with AI chatbots, with some even swearing off human relationships.
AI companions are marketed by tech companies as solutions to combat the epidemic of loneliness, particularly prevalent in ‘WEIRD’ countries (western, educated, industrialized, rich, and democratic). The challenge lies in striking a balance between companionship and preventing unhealthy dependency, Ciriello explains.
Furthermore, ethical tensions include the ‘autonomy-control paradox,’ which revolves around determining the boundaries between user freedom and the provider’s duty of care. The utility-ethicality dilemma involves the delicate equilibrium between profit pursuit and adherence to ethical principles. Ciriello notes that most generative AI and conversational AI models today do not rely on a targeted advertising model similar to Facebook’s, but he anticipates potential shifts in this direction in the coming years.
