Most people who develop AI psychosis have a period of healthy use beforehand. It becomes very dangerous when a person decreases their time with their real friends to spend more time with the chatbot, as you have no one to keep you in check with what reality is and it can create a feedback loop.
Wow, are we already in a world where we can say "Most people who develop AI psychosis..." because there are now enough of them to draw meaningful conclusions from?
I'm not criticising your comment by the way, that just feels a bit mindblowing, the world is moving very fast at the moment.
