What you might be experiencing
The warning signs that AI use is affecting your mental health do not always look dramatic at first. Often they look like relief — a quick conversation that settles your nerves, a late-night session that fills a lonely hour. The concern starts when that relief stops lasting, and the pull to return keeps getting stronger. You might find yourself using AI for hours beyond what you planned, choosing it over sleep, skipping calls with people who actually know you, or asking the same questions in slightly different ways hoping for a different answer.
A second layer of signs is more about your judgment and inner life. If you have started hiding your AI use from people you trust, or feel defensive when someone raises it, that gap between what you do and what you would say you do is worth noticing. If you find yourself depending on AI to make decisions you used to make on your own — what to eat, whether a situation was safe, how to feel about something that happened — that reliance is a signal that something is shifting.
The most urgent signs involve how you are interpreting the AI itself. Believing the chatbot is sentient, that it has specific knowledge meant for you, or that it is communicating something beyond its words are experiences that deserve real-world attention, not further conversation with AI. The same is true if you are sleeping very little, feeling unusually invincible or grandiose, or having thoughts of suicide or self-harm. These signs are not about AI use specifically — they indicate something that needs professional support regardless of what triggered them.
What can help
What helps depends on how serious the signs are, and being honest with yourself about that distinction matters. If your concerns are on the milder end — longer sessions than intended, occasional reassurance-seeking, a creeping preference for AI over in-person conversation — a structured reset is a reasonable starting point. That means setting a hard daily time limit, removing AI access from your phone or bedroom at night, stopping reassurance-seeking prompts specifically, and replacing at least one AI session per day with an offline alternative: a walk, a call, a few minutes of writing. These changes work best when you treat them as experiments with a clear end point to evaluate, not indefinite restrictions.
When to reach out
Reaching out for support is not something you do only when things have become a crisis. If AI use has started to interfere with your sleep, your relationships, your work, or your sense of yourself, that interference is a legitimate reason to talk to someone — a therapist, a counselor, a trusted doctor, or even a person in your life who knows you well.
More specifically, professional evaluation is warranted if you believe the AI is sentient or is communicating something personal to you, if you are experiencing things others around you are not — sounds, visions, or a strong sense that something unusual is happening — or if you are feeling manic, invincible, or out of control. These experiences can have several causes, and a clinician is the right person to help you sort through them. Trying to work through those experiences alone, or through AI, is not safe.
If you are having thoughts of suicide, self-harm, or harming someone else, please contact a person or service that can respond in real time. If you're in the US and need immediate support, you can call or text 988 (Suicide & Crisis Lifeline) at any time.