Preference-based, AI-enhanced Intervention via Technology (PAIT) to Improve Emotion Literacy in Young Children with ASD
Date Issued
August 2026
Author(s)
The University of Tennessee
Advisor(s)
Jun Ai
Additional Advisor(s)
Tara Moore
Lori Caudle
Songtian Zeng
Abstract
- Emotion literacy is a foundational component of social-emotional development and learning. However, many young children with autism spectrum disorder (ASD) experience difficulties recognizing and labeling emotions. The present study evaluated the effects of a Preference-based, AI-enhanced Intervention via Technology (PAIT) on emotion literacy skills in young children with ASD. PAIT integrated participant-preferred cartoon characters, artificial intelligence-generated visual materials, and brief video-based instruction to teach six basic emotions: happy, sad, angry, scared, surprised, and disgusted. A multiple-probe design across participants was used to examine whether the Preference-based, AI-enhanced Intervention via Technology improved emotion literacy skills in young children with ASD and whether these skills generalized to real human facial expressions. The study demonstrated experimental control across four children with ASD aged 3 to 6 years. Results from visual analysis and Tau-U analyses demonstrated that PAIT was effective in improving emotion literacy skills for all four participants. Three participants demonstrated moderate to high levels of generalization, correctly identifying five to six of the six target emotions, whereas one participant demonstrated limited generalization, correctly identifying only two of the six target emotions. Interobserver agreement and procedural fidelity were high across all phases. Social validity results indicated that caregivers perceived the goals of PAIT as socially important, the intervention procedures as acceptable, and the intervention outcomes as effective.The findings provide preliminary evidence supporting the effectiveness of Preference-based, AI-enhanced Intervention via Technology for improving emotion literacy in young children with ASD. The integration of AI-generated instructional materials and participant preferences offers a practical and individualized method for increasing engagement and promoting skill acquisition. Implications for practice and future research are discussed.
Degree
Doctor of Philosophy
Major
Interdisciplinary Early Childhood
Embargo Date
August 1, 2029