Effective for What Kind of Students? The Changes Brought to Personalized Learning Environments by Interactive Generative AI
SCC Co., Ltd. announced joint research results with Professor Kitazawa of Aoyama Gakuin University on their AI review support service 'Manaripi'. A trial involving 5th-grade students confirmed positive behavioral changes and a 3.3-point increase in standard deviation scores.
📋 Article Processing Timeline
- 📰 Published: April 15, 2026 at 00:00
- 🔍 Collected: April 14, 2026 at 15:31
- 🤖 AI Analyzed: April 19, 2026 at 16:05 (120h 34m after Collected)
SCC Co., Ltd. (Headquarters: Nakano-ku, Tokyo; Representative: Kunihiko Kasuga; hereinafter SCC) announced that the results of joint research on the 'AI Review Support Service "Manaripi"' (hereinafter 'Manaripi') with Professor Takeshi Kitazawa of the Department of Education, College of Education, Psychology and Human Studies at Aoyama Gakuin University (Headquarters: Shibuya-ku, Tokyo) were presented at the 'Educational Society in the AI Era 2025 2nd Workshop' held on Sunday, March 22, 2026.
Presentation Information
Educational Society in the AI Era Official Website: https://sites.google.com/eduaiera.org/home/home
Overview and Background of the Joint Research
The purpose of this joint research was to develop 'Manaripi' and verify its effectiveness in personalized learning environments. This initiative came to fruition because Professor Kitazawa resonated with SCC's desire to 'realize a society where all people living in the future information society can live prosperously.' During the presentation at the 'Educational Society in the AI Era 2025 2nd Workshop,' insights gained through 'the development and proof-of-concept experiments of Manaripi' in the joint research thus far were highlighted.
Research Details and Results
'Manaripi' was freely used by 5th-grade students (72 individuals) at a private elementary school in the Kanto area for reviewing history outside of standard class hours.
Main Features of Manaripi:
After the AI reads educational materials and lesson videos created by teachers, it generates review data using proprietary technology to support students' personalized learning through the following functions:
- Extracts important knowledge and poses questions in the form of an avatar speaking to the user
- Users explain their answers to the AI avatar in their own words
- Real-time assistance for mistakes or stumbling blocks
- Supports providing hints, multiple-choice questions, and repeated questioning
(*As of April 1, 2026)
The demonstration period was from Saturday, November 15, 2025, to Wednesday, November 26, 2025. In addition to 67 valid responses obtained from a questionnaire survey, analysis and evaluation were also conducted based on learning logs and mini-test results (standard deviation scores).
The questionnaire survey revealed positive responses regarding the perception of 'Manaripi', with significant differences recognized in 6 out of 8 items.
Items where significant differences were recognized:
- Reviewing becomes fun
- I can understand what I studied in social studies
- I know where I am stumbling
- Easier to understand compared to usual reviewing
- I can study at my preferred time
- I want to recommend it to other elementary school students
A particularly notable point is that a remarkable change was observed in students who continuously used 'Manaripi' (8.9%). These students previously showed tendencies such as 'not knowing how to study' and 'finding it hard to ask questions in class.' However, analysis of the learning logs showed actions such as 'asking questions spontaneously,' 'trying again based on hints,' and 'leading to continuous learning' through the use of 'Manaripi.' Furthermore, it was confirmed that these specific students improved their standard deviation scores by an average of 3.3 in mini-tests.
Future Developments and Prospects
From this joint research, it was observed that 'Manaripi' tends to be particularly effective for 'students who do not know how to study' and 'students who are not good at asking questions during class.' It became clear that 'Manaripi' has a high potential to be an effective approach to grasping and following up on the understanding level of students whose learning comprehension and stumbling blocks are difficult to surface. Moving forward, we plan to upgrade the system with functions that help grasp students' understanding of class content, making it even more useful for teachers in educational settings.
Comment from Professor Kitazawa
Through this research, we found that interactive generative AI can be expected to be utilized as a means for students to recognize where they are stumbling when they get stuck while learning. Moving forward, we will continue conducting ongoing verifications while proceeding with analysis and development from the perspective of utilizing educational data so that teachers and learners can grasp their level of understanding.
Scene from the Presentation (Professor Kitazawa, Aoyama Gakuin University)
About Manaripi
Born from joint research with Aoyama Gakuin University, 'Manaripi' is an unprecedented AI review support service that encourages students to 'explain what they have learned in their own words' through dialogue with an AI partner on a learning device. As the AI partner poses optimal questions tailored to each student's level of understanding, it goes beyond passive knowledge input, creating opportunities for output where students organize their thoughts and 'explain in their own words.' Through this, within learning...
Presentation Information
Educational Society in the AI Era Official Website: https://sites.google.com/eduaiera.org/home/home
Overview and Background of the Joint Research
The purpose of this joint research was to develop 'Manaripi' and verify its effectiveness in personalized learning environments. This initiative came to fruition because Professor Kitazawa resonated with SCC's desire to 'realize a society where all people living in the future information society can live prosperously.' During the presentation at the 'Educational Society in the AI Era 2025 2nd Workshop,' insights gained through 'the development and proof-of-concept experiments of Manaripi' in the joint research thus far were highlighted.
Research Details and Results
'Manaripi' was freely used by 5th-grade students (72 individuals) at a private elementary school in the Kanto area for reviewing history outside of standard class hours.
Main Features of Manaripi:
After the AI reads educational materials and lesson videos created by teachers, it generates review data using proprietary technology to support students' personalized learning through the following functions:
- Extracts important knowledge and poses questions in the form of an avatar speaking to the user
- Users explain their answers to the AI avatar in their own words
- Real-time assistance for mistakes or stumbling blocks
- Supports providing hints, multiple-choice questions, and repeated questioning
(*As of April 1, 2026)
The demonstration period was from Saturday, November 15, 2025, to Wednesday, November 26, 2025. In addition to 67 valid responses obtained from a questionnaire survey, analysis and evaluation were also conducted based on learning logs and mini-test results (standard deviation scores).
The questionnaire survey revealed positive responses regarding the perception of 'Manaripi', with significant differences recognized in 6 out of 8 items.
Items where significant differences were recognized:
- Reviewing becomes fun
- I can understand what I studied in social studies
- I know where I am stumbling
- Easier to understand compared to usual reviewing
- I can study at my preferred time
- I want to recommend it to other elementary school students
A particularly notable point is that a remarkable change was observed in students who continuously used 'Manaripi' (8.9%). These students previously showed tendencies such as 'not knowing how to study' and 'finding it hard to ask questions in class.' However, analysis of the learning logs showed actions such as 'asking questions spontaneously,' 'trying again based on hints,' and 'leading to continuous learning' through the use of 'Manaripi.' Furthermore, it was confirmed that these specific students improved their standard deviation scores by an average of 3.3 in mini-tests.
Future Developments and Prospects
From this joint research, it was observed that 'Manaripi' tends to be particularly effective for 'students who do not know how to study' and 'students who are not good at asking questions during class.' It became clear that 'Manaripi' has a high potential to be an effective approach to grasping and following up on the understanding level of students whose learning comprehension and stumbling blocks are difficult to surface. Moving forward, we plan to upgrade the system with functions that help grasp students' understanding of class content, making it even more useful for teachers in educational settings.
Comment from Professor Kitazawa
Through this research, we found that interactive generative AI can be expected to be utilized as a means for students to recognize where they are stumbling when they get stuck while learning. Moving forward, we will continue conducting ongoing verifications while proceeding with analysis and development from the perspective of utilizing educational data so that teachers and learners can grasp their level of understanding.
Scene from the Presentation (Professor Kitazawa, Aoyama Gakuin University)
About Manaripi
Born from joint research with Aoyama Gakuin University, 'Manaripi' is an unprecedented AI review support service that encourages students to 'explain what they have learned in their own words' through dialogue with an AI partner on a learning device. As the AI partner poses optimal questions tailored to each student's level of understanding, it goes beyond passive knowledge input, creating opportunities for output where students organize their thoughts and 'explain in their own words.' Through this, within learning...