Large language models are enabling NPCs that can hold genuine conversations, remember past interactions, and exhibit personalities that feel startlingly human. This is exciting for game design, but it raises important questions.
What happens when players form emotional attachments to AI characters? How do we handle scenarios where NPCs might be subjected to virtual harm? Should AI characters have some form of simulated consent?
These are not just philosophical exercises - they are design decisions we face daily. We share our framework for ethical AI character design and invite the community to join this important conversation.