This study develops and validates the Chatbot-Assisted Developmental Assessment Scale for Preschool Children (CADAS), a screening tool for 1-6-year-olds. Using AI and the Kebbi Air S robot, CADAS assesses motor, cognitive, language, and social-emotional development through interactive games. A total of 180 children will participate, comparing AI-robot and human assessments for reliability and validity. The project aims to create a localized, efficient, and accurate developmental screening tool to support early detection and improve child development services in Taiwan.

Current developmental assessments in Taiwan rely mainly on foreign tools, which limit cultural suitability. This study addresses that gap by creating a local, AI-based screening system using Kebbi Air S, a social robot capable of voice prompts, facial recognition, and data recording. Five game-like scenarios will assess different developmental domains; one will be robot-assisted and four human-administered for comparison. 180 children (1-6 years) will be recruited from a nonprofit preschool in Taoyuan, with 60 also taking the Comprehensive Developmental Inventory for Infants and Toddlers (CDIIT) for validation. Analyses using SPSS 21.0 and ConQuest 4.5 will test reliability, validity, and item quality. The study aims to build a reliable local tool and confirm the feasibility of chatbot-assisted evaluation, contributing to a national preschool development database and improving early intervention accuracy.

Inclusion Criteria:

• Children aged 1-6 years enrolled in preschools or nurseries in Taoyuan City.
• Parent or guardian consent obtained.
• Able to follow basic instructions and complete play-based tasks

Exclusion Criteria:

• Diagnosed developmental, neurological, or genetic disorders.
• Severe visual, hearing, or motor impairments affecting task performance.
• No parental consent or early withdrawal during assessment.

Preschool children from Taoyuan City, Taiwan.