Early detection of children's hidden signals by AI Diary

Implemented created by Masae Ebersson English
73d ago update

In 2020, Japan recorded an alarming increase in student suicides, with 415 cases among schoolchildren aged 6 to 18. 
It is 31% higher than the previous year and the highest since records began in 1974. 

This surge highlighted the urgent need for early detection and support for children's mental health, particularly as the COVID-19 pandemic disrupted traditional school and family environments. 

In response, Katsuragi city in Nara, Japan, launched the first AI Diary project in April 2022. This initiative enables students to write diaries anonymously on educational tablets, which are then analyzed by AI to identify emotional distress and provide timely support.

This system is primarily created for junior high school students (ages between 12-16), however, there are lots of requests to expand to additional schools  (ages between 8 - 16), which is around 30,000 elementary and junior high schools in Japan.

Objectives:

The primary goal of the AI Diary and chat system is to establish a proactive approach to monitoring student mental health by analizing their diaries (anonymous) on educational tablets.

With regard to human-centred technology, the school teachers had many workshops and meetings with AI business developers to frame the problem and create solutions that match their students’ needs. They wanted to invent not only an AI diary but also a chat system on the same tablet, so that students can also interact with registered social service staff /therapists on chat conversations. This is accessible anytime of the day. 

Combining AI Diary and a chat system with real human therapists helps students to express themselves in different occasions and styles. 

By leveraging technology to recognize early signs of anxiety and distress, the project aims to prevent suicides and foster a supportive environment where children feel safe to express their concerns.

Solution:

The AI Diary operates by allocating 10 minutes each Friday for students to write in their diaries anonymously. 

AI analyzes the language and context of these entries to detect strong emotions or repeated themes. If the system identifies a student in need of support, it automatically sends a message encouraging them to seek help. 

Data was trained by Emergency Response Guidance and Abuse Response Guide for Schools/Board of Education, as well as a special word dictionary of people’s diary who have attempted suicide and those who considering it.  Emotional analysis is performed by Google API.

Natural language processing system MIS engine is set between a diary page for children and a domain for administrators. The contents of the forms filled out by the children will be sent to the MIS engine and encrypted, thus the children's privacy will be protected.

However, in case of a life-threatening emergency, the school can override Japan's Personal Information Protection Law to identify the student. Additionally, a chat feature allows students to interact with registered social service staff or therapists, providing 24/7 access to mental health support. 

This dual approach of combining AI analysis with real human interaction enhances the effectiveness of the support system.

Outcome:

Since its launch, the AI Diary system has successfully facilitated over 10,000 student interactions. It has proven to be an invaluable tool for early intervention, allowing educators to respond effectively to students' emotional needs. 

In addition to offering advice, the system encourages students to connect with adults by asking questions like, "Is there an adult you can talk to?" or "Have you tried talking to a teacher?"

The system has garnered positive feedback from teachers, parents, and students, fostering a greater sense of security and openness. As the project evolves, there are commercial opportunities to expand this initiative to other schools in Japan, targeting approximately 30,000 elementary and junior high schools. 

By addressing mental health proactively, the AI Diary project stands as a model for integrating technology in education and safeguarding children's well-being.

Attributes

Education
Child Protection
Better Customer Experience, Smarter Product or Service
Language
NLP