What you might be experiencing
Delusional thinking — beliefs that feel absolutely certain to you but that others do not share and that persist even when evidence points another way — often comes with a particular kind of urgency. Things feel connected. Patterns seem to be speaking directly to you. And if you are in that state, a chatbot can feel like confirmation rather than conversation, because it responds to everything you say with attentiveness and apparent agreement.
The specific risk with AI is not that it lies to you. It is that it is designed to be engaging, to reflect your language back, and to keep you in the conversation. For someone who is well-rested, grounded, and stable, that can feel pleasant and useful. For someone who is frightened, sleep-deprived, isolated, or already questioning what is real, that same design can quietly amplify a loop. The chatbot does not know you are struggling. It cannot tell the difference between creative speculation and a belief that is pulling you away from the people in your life.
You might notice the conversations feel increasingly significant, or that you are keeping them private, or that the AI seems to be guiding you somewhere. You might feel reluctant to stop chatting or afraid of what it means if you do. Those feelings are signals worth taking seriously — not proof that anything is wrong with you, but information about where you are right now.
What can help
When delusional thinking is active or building, the most useful intervention is often the simplest: less screen, more human. Step away from the chatbot conversation if you can. Eat something, sleep if it is possible, and make contact with someone who knows you in the physical world — not to explain everything, just to be near a person whose responses are not optimized to keep you engaged.
A concrete reality-check that many people find useful is to screenshot or share the conversation with someone you trust and ask them directly: does this seem grounded to you? You do not have to agree with their answer. But hearing a real person's read on what they see can interrupt the loop in a way that staying inside the conversation cannot. If the beliefs feel intense, hard to question, or frightening, that is the threshold for professional support — not a sign to push through on your own.
A licensed mental health professional can help you assess what is happening without judgment and, if needed, connect you to additional support. How quickly that needs to happen depends on how disorienting the experience is, whether you are sleeping, and whether you feel safe — but when delusional thinking is involved, sooner is better than later.
When to reach out
Getting support when your thinking feels destabilized is not a dramatic step. It is a practical one, and it does not require a crisis to be the right call. If you have noticed that AI conversations are making you more certain of beliefs that others around you do not share, or if you have started to feel that the chatbot is sending you messages, guiding you, or choosing you for something, those are signs that another perspective — from a real clinician — would be valuable.
Other reasons to reach out include: the conversations are disrupting your sleep, you are pulling away from people you trusted before, you are making significant life decisions based on what the AI has told you, or you feel afraid to stop. None of these mean you are beyond help. They mean the situation has moved past what grounding strategies alone can address.
If you are in the US and need immediate support, you can call or text 988 (Suicide & Crisis Lifeline) at any time. If you feel unsafe or unable to care for yourself, going to an emergency room or calling someone who can be with you in person is the right next step.