By analyzing human-AI interaction through the lens of creative pretense and media theory, the study explains why people readily form emotional bonds and confide in AI companions without attributing moral status to them, allowing users to feel affection while freely resetting memories or switching off systems.