OpenAI is worried that ChatGPT-4o users are developing feelings for the chatbot

Like the previous commenter, I'm not convinced that OpenAI is worried about this. Anthropomorphizing AI and computers in general has been a problem since before computers existed. It's a trope that goes all the way back to sci-fi literature.

Play-pretending that AI is a human isn't any more dangerous than any other kind of play. Even a game like My Time at Sandrock has character dialog in which both the human and NPC "pretend" to be in a relationship. No one is worried about that.

This reads more like ad copy from OpenAI than anything reality related.