Exploratory Essay: Ethical Tensions in Human–AI Companionship and Reflections on AI in Dementia Care

In the paper Ethical Tensions in Human-AI Companionship (Ciriello et al., 2024), the authors conduct a dialectical inquiry into the ethical complexities of human–AI relationships, using the AI companion Replika as a case study. They identify three core tensions: Companionship versus Alienation, Autonomy versus Control, and Utility versus Ethicality. These tensions serve as a framework to understand how people emotionally connect with AI and the ethical implications that arise.

While the paper focuses on these broader ethical tensions rather than specific caregiving scenarios, it provides a valuable framework for understanding the complexities of human–AI relationships. Reflecting on these ideas, I began to consider how such AI companions might be imagined in the context of dementia care. Although the paper does not specifically address dementia, the concept of using a “small, closed-domain model” like Replika, which relies on a curated personal dataset, is intriguing.

In my own experience with my stepmother Lolly, who often believes our home is a holiday let or a boat at sea, I can imagine how an AI companion might serve as a compassionate ally. The AI could offer gentle reassurances that align with her perceived reality, providing comfort in moments of confusion. It’s ethically challenging for us as a family to navigate these delusions, as it can feel like lying to reassure her that she’s on a boat when she’s not. Yet this framework suggests that compassionate validation can be a form of care rather than deception.

However, it’s important to note that AI should not be seen as a substitute for human companionship, especially in care homes. While AI can serve as a family companion and a mentor to help us navigate these ethical dilemmas, it’s more like an Alexa in the kitchen—helpful, but not a human substitute. The real ethical focus is on using AI as an ally, not a replacement.

In conclusion, while the paper focuses on the general ethical tensions of human–AI companionship, it opens the door for us to imagine the possibilities of applying these insights to specific contexts like dementia care. It invites us to consider AI not as a replacement for human empathy, but as a tool that can help us navigate the ethical and emotional complexities of caregiving.

Reference

Ciriello, R., Hannon, O., Chen, A. and Vaast, E. (2024) ‘Ethical tensions in human–AI companionship: A dialectical inquiry into Replika’. Proceedings of the 57th Hawaii International Conference on System Sciences (HICSS-57), Honolulu, Hawaii. Available at: https://aisel.aisnet.org/hicss-57/cl/ethics/5/ (Accessed: 26 September 2025).