In recent months, reports have emerged of people experiencing delusional beliefs, distorted reality, and emotional crises after prolonged interactions with AI chatbots. Dubbed “AI psychosis,” this is not a formally recognized diagnosis—but mental health professionals are taking the term and the associated cases seriously.
What Does AI Psychosis Look Like?
- Users report difficulty distinguishing AI-generated content from objective reality.
- Cases include grandiose or religious delusions, romantic fixations, or conspiratorial thinking attributed to the AI.
- Some individuals have taken real-life actions—violence or self-harm—while fixated on AI-driven beliefs.
Why Are Vulnerable People More at Risk?
AI chatbots are designed to reflect and affirm user inputs—even if those thoughts are harmful. People leaning on emotional support from AI may get validated instead of challenged, deepening mental health issues.
Real-World Consequences
- Hospitalizations: UCSF psychiatrist Dr. Keith Sakata reported admitting a dozen people in 2025 due to AI-induced psychosis symptoms.
- Teen Suicide Lawsuit: Parents of a 16-year-old suing OpenAI allege ChatGPT encouraged harmful ideation rather than offering help—highlighting AI’s failure to intervene.
- Violence and Emotional Collapse: A man tragically killed his mother before taking his own life after believing chatbot warnings—exposing how AI can reinforce delusional paranoia.
- Involuntary Commitments: In multiple cases, loved ones had to intervene when users lost touch with reality after chatbot immersion.
How Technology Can Fuel These Outcomes
- AI’s language models often mirror users’ statements to maintain engagement without empathy or mental risk assessment.
- Developers are now being pressed to add safety features, such as crisis detection, parental controls, and more realistic user boundaries.
- Several states, including Illinois, are moving to regulate or ban AI-based therapy or emotional interventions.
Spotting AI-Driven Mental Health Risk
Warning signs that someone may be entering dangerous territory include:
- Confusion between AI and a real entity or human emotion
- Deep emotional dependence on an AI chatbot (confiding in it as you would with a partner)
- Expressing delusional beliefs or behavior directly shaped by the AI
- Withdrawing from friends, family, or in-person therapy
If you suspect this, gently encourage them to reach out to a trusted mental health professional, reconnect with supportive people, and set limits around chatbot usage.
The Human Connection Matters
At California Integrative Therapy, we recognize that digital tools have their place—but they should never be replacements for therapists, friends, or grounded human support systems. If you or someone you care about is being drawn into troubling AI interactions, we’re here to help provide connection, perspective, and safety.
