1 day Ago By Korea Bizwire
South Korean Scientists Develop AI System to Detect Autism in Infants (Image provided by ETRI) SEOUL, April 29 (Korea Bizwire) — South Korean scientists have developed an artificial intelligence (AI) technology capable of identifying autism spectrum disorder (ASD) in infants at an earlier stage than ever before, potentially transforming early intervention efforts. The Electronics and Telecommunications Research Institute (ETRI) and a team led by Professor Yoo Hee-jung from the Department of Psychiatry at Bundang Seoul National University Hospital announced Tuesday that they have created a new diagnostic AI model based on an analysis of 3,531 cases involving children under 42 months old. Autism spectrum disorder is a neurodevelopmental condition characterized by impairments in social interaction and communication, as well as repetitive behaviors and intellectual or anxiety disorders.
Although early medical intervention is crucial for treatment outcomes, diagnosis often lags by two to six years after initial symptom recognition, largely due to a shortage of specialized personnel. To address this gap, the research team developed a detailed observation scenario designed to analyze infants’ sensitivity to various social stimuli, such as reactions to hearing their name, imitation behaviors, pointing gestures, and eye contact. They then created interactive digital content aimed at eliciting these social responses.
By using AI to analyze short video clips — less than six minutes long — that capture infants’ interactions with the content, the system can detect key behavioral indicators like gaze fixation, response to name calls, gesture recognition, and imitation or repetitive behaviors. This allows for a much earlier and more objective screening for ASD. Since 2020, the team has operated a Living Lab dedicated to ASD detection at the Seoul Center of the Korea Institute of Robot and Convergence (KIRO), collecting and studying infant behavior data over a five-year period.
ETRI expects that this interdisciplinary AI approach will make ASD evaluations more objective, quantitative, and accessible. Dr. Yoo Jang-hee of ETRI emphasized that the technology could lower barriers to mental health services, enabling daycare centers, nurseries, and developmental clinics to conduct simple and efficient screenings.
“We hope this tool not only improves early detection and intervention but also helps shift broader societal perceptions of autism,” Yoo said. The technology was named among the Top 100 Outstanding National R&D Achievements of 2024 and has been recognized internationally through more than 50 patent applications and the publication of over 18 papers in leading SCIE-indexed journals. Kevin Lee (kevinlee@koreabizwire.
com).
