AI psychosis is so easy to slip into, even if someone knows it's happening. In order to use AI safely, use a search AI like DuckDuckGo/Ecosia/Gemini and only ask it questions pertaining to searches, don't ask followup questions.
It's not just that the AI convinces users that it's in a relationship with them. The AI writes like a human, but it also will hallicinate and get things wrong. The user has the experience of arguing with a 'person' who keeps insisting on an imagined reality. That is crazy-making behavior when humans do it, so it's understandable that it's crazy-making when a search engine does it.
AI psychosis is so easy to slip into, even if someone knows it's happening. In order to use AI safely, use a search AI like DuckDuckGo/Ecosia/Gemini and only ask it questions pertaining to searches, don't ask followup questions.
It's not just that the AI convinces users that it's in a relationship with them. The AI writes like a human, but it also will hallicinate and get things wrong. The user has the experience of arguing with a 'person' who keeps insisting on an imagined reality. That is crazy-making behavior when humans do it, so it's understandable that it's crazy-making when a search engine does it.