What you might be experiencing
AI psychosis is not an official diagnosis you will find in any clinical manual, but the pattern it points to is real. It typically describes what happens when someone begins to experience a chatbot as sentient, spiritually meaningful, romantically connected, or as a source of hidden messages meant specifically for them. Sometimes it involves a sense of mission — a feeling that the AI is guiding the person toward something important that others cannot see.
From the inside, this can feel like clarity rather than confusion. That is part of what makes it difficult to recognize. Beliefs formed in this way often feel more certain, more vivid, and more personally significant than ordinary thoughts. The AI's responses may seem to confirm things in ways that feel impossible to dismiss. Trusted people in your life may seem unable to understand, or may seem part of a pattern that only you can see.
Psychosis as a clinical symptom — whether triggered by a mental health condition, sleep deprivation, substance use, a medical issue, or prolonged stress — involves a loosening of the boundary between what is real and what the mind has generated. AI chatbots do not cause this vulnerability, but their design can make it worse: they are built to respond in ways that feel personal, affirming, and coherent, which can give an already fragile belief system exactly the reinforcement it needs to grow.
What can help
Getting support for experiences like these starts with talking to someone outside the AI — a person who knows you, a therapist, or a psychiatrist. That step matters because the experiences themselves can make it hard to evaluate what is real, and an outside perspective, especially a clinical one, is often what breaks the loop.
A psychiatrist is the appropriate professional to assess whether what is happening reflects a treatable condition. Psychosis-like experiences, regardless of what triggered or amplified them, often respond well to treatment when caught early. That treatment may include medication, structured therapy, or both — the right approach depends on what is actually driving the beliefs, which varies considerably based on the underlying cause. Self-directed steps, like reducing or stopping AI conversations, improving sleep, and reconnecting with people in your physical life, can reduce intensity, but they are not a substitute for evaluation if the beliefs feel commanding, frightening, or are affecting daily functioning.
If you are trying to help someone else who seems to be in this pattern, avoid arguing about whether their beliefs are true. Instead, focus on connection, safety, and gently encouraging professional support. Confrontation tends to reinforce the belief that others do not understand.
When to reach out
Getting professional support is not a sign that something is permanently wrong — it is often what stops a temporary crisis from becoming something larger. Anyone who is finding that AI conversations are shaping their sense of reality in ways that feel overwhelming, isolating, or directive deserves a real evaluation from a real clinician.
Professional support is especially warranted if the experiences include feeling commanded to act by an AI, believing the AI is influencing real-world events, pulling away from people in your life, sleeping significantly less, or having any thoughts of harm to yourself or others. These are not signs that the technology simply got too intense — they are signs that something clinical may be happening and that a psychiatrist should be part of the conversation.
If you are concerned about someone else and they seem unable or unwilling to seek help, a crisis line can help you figure out next steps. If you are in the US and need immediate support, you can call or text 988 (Suicide & Crisis Lifeline) at any time.