Skip to main content Skip to navigation

BrickSmart: Leveraging Generative AI to Support Children's Spatial Language Learning in Family Block Play

Project Overview

The document explores the application of generative AI in education, particularly through tools like BrickSmart, which enhances spatial language learning for children via guided block play. This AI-driven platform offers personalized building instructions and adaptive vocabulary support, fostering productive parent-child interactions. A comparative study demonstrated notable advancements in children's spatial vocabulary and increased parental confidence in their ability to support learning. These findings underscore the potential of generative AI not only to tailor educational experiences to individual needs but also to improve overall educational outcomes in early childhood settings. The use of such innovative technologies signifies a transformative shift in educational practices, emphasizing the benefits of personalized learning and parental engagement facilitated by AI.

Key Applications

BrickSmart

Context: Family block play for children aged 6-8 years

Implementation: BrickSmart uses generative AI to create personalized building instructions and vocabulary guidance, integrating a three-step process for guided play.

Outcomes: Significant improvements in children's spatial vocabulary use and parental capability in guiding learning.

Challenges: Parents may lack expertise in spatial language; ensuring effective interaction between AI guidance and parental involvement.

Implementation Barriers

Skill Gap

Many parents lack the expertise to effectively guide children's spatial language development during block play.

Proposed Solutions: BrickSmart provides structured guidance and prompts to empower parents in facilitating learning.

Cognitive Load and Parental Involvement

Parents may experience increased cognitive load when following detailed instructions from the system, which can also challenge meaningful parent-child interaction and reduce spontaneous exchanges.

Proposed Solutions: Future iterations of the system could incorporate multimodal aids (animations, videos) to alleviate cognitive demands, and include customizable guidance options for parents to adapt AI input based on their comfort.

Project Team

Yujia Liu

Researcher

Siyu Zha

Researcher

Yuewen Zhang

Researcher

Yanjin Wang

Researcher

Yangming Zhang

Researcher

Qi Xin

Researcher

Lunyiu Nie

Researcher

Chao Zhang

Researcher

Yingqing Xu

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Yujia Liu, Siyu Zha, Yuewen Zhang, Yanjin Wang, Yangming Zhang, Qi Xin, Lunyiu Nie, Chao Zhang, Yingqing Xu

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies