Researchers Develop AI That Detects Autism In Six Minutes

South Korean researchers have developed an AI system that can help detect autism spectrum disorder (ASD) signs in infants and toddlers, using just a six minute assessment video.

The new “social interaction recognition AI” was developed by the Electronics and Telecommunications Research Institute (ETRI) in collaboration with Seoul National University Bundang Hospital.

EurekaAlert explains that the system works by showing specially designed video content to young children, and analyzing their social responses using cameras and AI. These responses include eye contact, pointing, imitating actions, and reacting when their name is called.

Dr. Yoo Jang-Hee, Principal Researcher at ETRI, said: “We hope that this will help shorten the time between symptom detection and diagnosis, along with changing societal perceptions of autism. In addition, it is important for our research to solve hard problems, but we also hope that it will also contribute more to solving important problems like autism,” as reported by EurekaAlert.

Typically, medical professionals can identify ASD in children between 12 to 24 months of age. However, the researchers explain that due to the shortage of experts, and limited resources, the formal diagnosis are typically delayed. Previous research notes how early detection and support can significantly improve developmental outcomes.

The new AI screening system was tested with 3,500 children under 42 months old through a multidisciplinary, AI-based approach. EurekaAlert reports that this new approach represents the first system of its kind for autism detection.

The innovation received top 100 research achievement status in South Korea for 2024 and will become available in homes, daycares, and health centers to provide early support to children who might need this.