Nearly 1 in 5 US teens and young adults have sought mental health advice from AI chatbots, study finds

CLAIM (Claim-Specific Citation Network audit, sometimes called CSN) is a forensic method for testing whether a scientific or medical claim's authority is supported by evidence or by citation dynamics. It detects citation bias, amplification, citation diversion, citation transmutation, dead-end citation, and back-door invention.

The ASSAY skill runs a structured, CLAIM-compatible extraction and integrity assessment on an article. Output is a verdict (sound, mixed, flagged, problematic, or cascade), a count of claims extracted, the central key claim, and an integrity note describing the structural read.

This scan restricts ASSAY to peer-reviewed publications and preprint servers. Journalism, opinion pieces, and government documents are evaluated under different frameworks (CAIHL for power and agency; editor's note for context).

ASSAY found the central claims well-supported by the underlying evidence; methodology stands; the integrity-of-citation check raised no structural concerns.

19.2% of US adolescents and young adults ages 12-21 used AI chatbots for mental-health advice in November 2025, up from 13.1% one year earlier; the majority did not disclose this use to anyone.

Total claims extracted from the article: 9. The key claim is the single most load-bearing assertion the rest of the argument depends on.

Cross-sectional nationally representative RAND American Youth Panel survey (n=1,009), published in JAMA Pediatrics June 1 2026 (DOI 10.1001/jamapediatrics.2026.2015). Single time-point design with one prior comparator; the 13.1% → 19.2% trend is two cross-sections, not a panel. Self-report of stigmatized behavior risks under-reporting, which would make the 19.2% a lower bound. The disclosure-secrecy finding is the most actionable claim and the most exposed to social-desirability bias.

Editor's note: Prevalence number that should be in every Colorado HB26-1195 floor speech. The behavior the law is constraining is already the dominant behavior in the cohort the law is constraining — and most of them are hiding it.

Summary: Click2Houston: Coverage of McBain et al., JAMA Pediatrics, June 1 2026 — a nationally representative RAND survey (n=1,009 ages 12-21) finds 19.2% used AI chatbots (ChatGPT, Gemini, Character.AI, Meta AI) for mental-health advice, up from 13.1% a year earlier; a majority of those users kept the use secret.

Read the original source → CAIHL read of this item →