ARTIFICIAL INTELLIGENCE–ASSISTED EARLY AUTISM SCREENING: INTEGRATING M-CHAT-R/F WITH DIGITAL BEHAVIORAL MARKERS

Abstract

Autism-specific screening can reduce delays in recognition, but accuracy and access vary across settings. M-CHAT-R/F is a validated two-stage parent-report tool for ages 16–30 months; it estimates likelihood and does not establish diagnosis. AI can combine questionnaire responses with video, gaze, speech, or movement features, potentially improving triage while introducing risks of bias, privacy loss, and dataset shift [1–5].