Skip to main content Skip to navigation

From Keyboard to Chatbot: An AI-powered Integration Platform with Large-Language Models for Teaching Computational Thinking for Young Children

Project Overview

The document discusses an innovative AI integration platform named Spark, which aims to teach computational thinking to children aged 4-9 through an interactive conversational interface with a robot. By employing a hybrid pedagogy that merges top-down and bottom-up teaching strategies, Spark allows young learners to express programming tasks in natural language. The system then translates these tasks into executable code using a specialized Spark programming language, facilitated by a large language model (LLM) that enables semantic decomposition of tasks. This approach makes programming more accessible, eliminating the need for keyboard skills, and promotes deeper engagement with programming concepts. Through hands-on interactions with a tangible robot, the platform not only enhances understanding but also fosters enthusiasm for learning programming in early education. Overall, the findings highlight how generative AI can revolutionize teaching methodologies in education, making complex subjects like programming more approachable and enjoyable for young children.

Key Applications

Spark - AI-powered platform for teaching computational thinking

Context: Early childhood education (ages 4-9), specifically in programming and computational thinking skills

Implementation: Children interact with the system using natural language to describe tasks, which are processed by a chatbot and transformed into a Spark programming language code executed by a robot.

Outcomes: Improved engagement in learning programming concepts, reduced dependency on keyboarding, enhanced understanding of decomposition and sequencing in programming.

Challenges: Developmental readiness of young children, complexity of programming concepts, potential screen time concerns.

Implementation Barriers

Developmental Challenges

Young children may struggle with keyboarding and complex task decomposition, which are prerequisites for traditional programming education.

Proposed Solutions: Using natural language input and a conversational interface to bypass keyboarding requirements and simplifying the programming concepts.

Screen Time Concerns

Excessive screen time can be detrimental to young children's health.

Proposed Solutions: Integrating tangible interactions with a robot to minimize screen time while providing interactive learning experiences.

Project Team

Changjae Lee

Researcher

Jinjun Xiong

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Changjae Lee, Jinjun Xiong

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