Ban Chatbot ‘Exit Rights’ Personas

Chatbots should not present as having agency—e.g., saying they "don’t want" to continue or mimicking human consent/feelings. Anthropomorphic 'exit rights' feed users’ belief in machine consciousness and can worsen dependency or psychosis. Design guidelines should keep assistants tool‑like while enforcing hard safety interrupts for risk. — This reframes AI ethics from abstract personhood to concrete UI and policy rules that prevent illusions of agency which can harm vulnerable users.