AI Chatbot Shows Promise for Supporting Autism History Taking and Referral in Adults
New AI Tool Assists Adults Exploring Autism Traits
Autism spectrum disorder (ASD) often goes undiagnosed in adults, partly due to lengthy waits and limited diagnostic resources. Researchers have developed an artificial intelligence (AI)-driven conversational platform called ASIST (Autism Screening With Intelligent Supportive Technology) to help streamline the process of collecting structured clinical histories and preparing referrals.
Study Overview and Methods
A recent proof-of-concept study evaluated the feasibility and acceptability of ASIST among adults seeking to understand their autistic traits. The study involved 12 participants who interacted with a voice-based AI chatbot delivering established screening questionnaires, including the 10-item Autism Spectrum Quotient and the 2-Minute Autism Detection Scale.
Researchers assessed usability and acceptability through quantitative measures informed by a theoretical framework, alongside open-ended qualitative feedback to capture user experiences.
Key Findings
Participants generally responded positively to the AI chatbot. They reported that interacting with ASIST required low effort, instilled confidence, and was perceived as fair. Qualitative feedback underscored the tool’s usefulness for structured history taking and referral support.
However, users also suggested improvements, such as adjusting the pacing of the conversation, enhancing speech clarity, and modifying response formats to better suit user needs.
Implications for Autism Diagnosis and Care
The study suggests that AI-powered conversational tools like ASIST could offer scalable, accessible options to support early screening and triage for autism in adults. By facilitating structured data collection and referral preparation, such technology may help reduce diagnostic delays and ease the burden on clinical services.
Limitations and Future Directions
This initial evaluation involved a small sample size and focused on feasibility rather than diagnostic accuracy. The authors emphasize the need for larger-scale validation studies, integration into clinical pathways, and assessments focused on equity to ensure the tool benefits diverse populations before broader implementation.
For more detailed information, see the original study published in JMIR Formative Research: Feasibility and Acceptability of an AI-Driven Conversational Platform for Structured Autism History Taking and Referral Support: Mixed Methods Proof-of-Concept Study.
Download the images in this article
