Artificial Intelligence in Autism Assessment: A Conceptual Framework for Multimodal Artificial Intelligence Integration

Artificial intelligence (AI) may be increasingly positioned to strengthen autism spectrum disorder (ASD) assessment by improving scalability, consistency, and workflow efficiency across settings that serve children. Traditional diagnostic tools such as the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) and Autism Diagnostic Interview-Revised (ADI-R) require extensive clinician preparation and are difficult to implement in many contexts, which contributes to delays and uneven access. This article provides a detailed review of how AI methods can be embedded within existing evaluation procedures. Computer vision, speech analysis, natural language processing, predictive modeling, and multimodal fusion techniques are examined as tools that can support observation, interviewing, scoring, and documentation. A central contribution is a conceptual framework organized around structured signal extraction, interpretable and modular models, and strong safeguards that guide responsible use. Key challenges include bias, variation in data quality, and appropriate alignment with clinical judgment. The article offers a foundation for clinicians and researchers seeking to integrate AI within ASD assessment workflows.