Intelligent Educational Interventions and Demonstration Applications for Children with Autism

The UNESCO “Education 2030 Agenda” emphasizes the importance of learning that is inclusive, fair, and personalized. Following this vision, China’s Central Committee released Education Modernization 2035, which calls for advancing education modernization through informatization and stresses that children and adolescents with disabilities have the right to receive proper education.

In China, nearly 2% of children have autism spectrum disorder (ASD), and this number is rising quickly. More than ten million people are affected, most of whom cannot live independently. Autism is a serious social issue because it causes deep difficulties in social interaction and communication. There is no effective medical cure, and many individuals face lifelong disabilities. Therefore, education-based support is the main and most effective way to help. Meeting the educational needs of children with autism has become a major priority in China’s education development.

Compared to developed countries, China began autism rehabilitation efforts later. There is a shortage of trained professionals, serving fewer than 300,000 people. Interventions require intensive labor, which leads to high costs—averaging about 70,000 yuan per person per year. However, new advances in artificial intelligence are opening up opportunities for innovation in autism education. Early research abroad shows that intelligent education methods can help, but current practices are still basic, offering only simple personalization and shallow evaluation. These methods often miss the big differences between individual autistic children and fail to connect fragmented resources, rigid intervention models, and the complex developmental needs of these learners.

Three major technology challenges hold back progress in autism education:

  • Personalized Expression: Children with autism are very different from each other. Traditional methods do not capture changes in their thinking and emotions well, leading to poor personalization.

  • Resource Construction: Autistic learners have complex development paths that require multiple stages and different social learning scenarios. Current resources are scattered, inefficient, and not designed to fit their cognitive needs, often causing overload.

  • Process Intervention: Autism affects many areas—behavior, language, and emotions—and involves many factors. Existing interventions are often rigid, and commonly used tools like questionnaires and observations lack objectivity and accuracy.

To address these challenges, the National Engineering Research Center, together with Wuhan University’s School of Computer Science, Zhongnan Hospital, and Qihui Special Education School, has started a collaborative project. Supported by national and provincial programs—including the National Science and Technology Support Program, the National Key R&D Program, the National Natural Science Foundation, and others—this initiative focuses on using data sensing and behavioral analysis to overcome key barriers. It aims to model and interpret the cognitive and emotional states of autistic learners, build interactive human-computer learning environments, and develop adaptive, personalized educational interventions to meet each child’s unique needs.