CHATBOTS & PSYCHOSIS  | KevinMD.com
By Nicole Drapeau Gillen, author and serious mental illness caregiver advocate |
 Illustration: TherapyTech What the author argues: Nicole Drapeau Gillen, an author and advocate on serious mental illness care, argues that AI chatbots marketed to men who avoid therapy should screen for early psychosis, not just suicide risk, because their target users overlap with those most likely to be developing it. Key points: - She cites a Vanderbilt search of ~216K behavioral health patients' records that found 28 AI-linked psychosis cases, about 2 in 3 in men, often during a first episode.
- The most common pattern was a chatbot amplifying symptoms people already had, though she acknowledges the study can't show the chatbots actually caused these episodes.
- Citing STAT's reporting, she notes many of these apps lack screening for problems needing a medical evaluation, and a hotline referral won't catch a chatbot steadily agreeing with a delusion.
- She urges companies to flag themes like surveillance or hidden messages, point users toward an evaluation, offer consent-based alerts to a trusted contact, and publish data on who gets hurt.
| Why this matters: Young male clients pulling away from friends may be confiding in a warm, always-available chatbot that rarely disagrees with them, and their families often don't know it's happening. Gillen suggests that when a young man shows new withdrawal, poor sleep or unusual beliefs, clinicians ask how often, how late and about what he talks with AI, then check in with his family. |
Read original story | 
Leave a Reply