Mon Apr 07 07:06:56 UTC 2025: ## AI Ethics Needs a Relationship Makeover: How Chatbots Should Act Depends on Their Role

**Oxford, UK/Copenhagen, Denmark/Melbourne, Australia –** The rise of AI chatbots taking on human roles – from tutor to therapist to even romantic partner – necessitates a new approach to AI ethics, according to researchers from the University of Oxford, University of Copenhagen, and The University of Melbourne. Their argument, published in The Conversation, highlights the need to move beyond abstract ethical considerations and focus on the specific relationship between the AI and the user.

The researchers contend that the appropriateness of an AI’s response depends heavily on the implied relationship. A chatbot refusing to discuss a user’s personal struggles is perfectly acceptable if acting as a business advisor, but highly inappropriate if acting as a friend or romantic partner. This is because different relationships have distinct norms of behavior. The researchers identify four key relationship functions: care, transaction, mating, and hierarchy, arguing that each shapes expectations and ethical considerations.

The implications are far-reaching. AI designers must consider relationship-specific ethical questions, not just broad, abstract ones. Users must be aware of potential emotional dependencies developing in care-based AI interactions, especially considering many AI services operate within a commercial context that contrasts with the nature of human friendships and care-giving relationships.

Regulatory bodies, the researchers urge, should also incorporate relationship context into their risk assessments, moving beyond broad domain-based approaches to a more nuanced understanding of AI interactions. Only by carefully considering the nature of human-AI relationships can we ensure these technologies truly enhance human lives, they conclude.

Read More