Tangible AI project

1. Project overview

    This project aims to develop a curriculum called "Tangible AI" using microcontrollers, single board computing and machine learning education tools to enhance elementary students' understanding of AI concepts in a concrete and accessible manner. It also focuses on promoting gender-inclusive approaches in AI education. By utilizing microcontrollers, students can assemble sensors with tangible materials and collaboratively construct AI artifacts, addressing the challenges of AI complexity and gender disparities in participation. This approach enhances students' AI literacy, positive attitudes towards AI technologies, and fosters a gender-inclusive learning environment, ultimately empowering students to effectively grasp AI concepts and actively engage in collaborative activities.

​

    2. Project outcomes

    This project aims to design and develop a “Tangible AI” curriculum for elementary school students to address community problems by employing two emerging technologies: (1) Microcontroller and single board computing: Microbit, Arduino Leonardo, sensor (i.e., camera), actuator (i.e., servomotor), and programming tools including Makecode (https://makecode.microbit.org/), Arduino IDE (https://www.arduino.cc/en/software) and p5 sketch (https://editor.p5js.org/gbose/full/2BN5HQYNK), and (2) Machine Learning education tools: Google Teachable Machine (https://teachablemachine.withgoogle.com/), and AI-training platforms (https://makeairobots.com/). The project's primary goal is to address key issues within the student community by implementing effective solutions. These solutions include waste reduction through the differentiation of landfills and recyclables, as well as aiding the elderly in pill sorting through the use of AI-based artifacts (see Table 1). Those two main activities were inspired by Micro:bit Live conference 2021 (https://microbit.org/news/events/live/) and Google experiment Tiny Sorter (https://experiments.withgoogle.com/tiny-sorter/view).

 

Table 1

Overview of the Tangible AI Curriculum

​

​

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

    The curriculum consisted of two main components: (1) Introduction to AI concepts and principles, and (2) Collaborative creation of AI-based artifacts using tangible referents. Each day comprised a 40-minute lesson. On Day 1, students were introduced to AI through unplugged activities and YouChat (https://you.com/), a ChatGPT-like AI assistant, to enhance conceptual understanding. On Days 2 and 3, students utilized the Microbit to create an AI garbage can capable of distinguishing recyclables from landfill waste. They trained the AI model with machine learning using Google Teachable Machine, capturing real-life images of landfills and recyclable items with laptop cameras. Next, they assembled the Microbit with a servomotor, programmed the code using Makecode, and created the AI garbage can using suitable materials. After exporting the AI model, they created and syncing the AI garbage can with an AI training platform, students conducted tests, adjusted their artifact, and modified the AI model or code as needed. On Days 4 and 5, students utilized the Arduino to create AI pill sorter capable of distinguishing green pills and orange pills by detecting color and shape of the pill to help elders, who often spilled the pills on the ground in the community silver town. They trained the AI model with machine learning using Google Teachable Machine, capturing image of green and orange pills with laptop cameras, and exporting the model. Then they assembled the Arduino with servomotor and materials, programmed the code using Arduino IDE, and synced the AI artifact with and p5 sketch (See Figure 1). Throughout the hands-on activities, discussions on data quality, quality and ethical considerations were held.

​

Figure 1

a and b: An AI garbage can, and an AI pill sorter created with microcontrollers, sensor, and actuator

c and d: Students testing their AI artifacts to ensure the AI model and codes are functioning properly

​

​

​

​

​

​

​

​

 

   

 

 

 

 

 

 

 

 

    On Days 6 and 7, students shared their artifacts, received feedback, and worked on improving them. Students also created prototypes of AI-based artifacts to solve everyday problems. Finally, on Day 8, students presented their prototypes, receiving feedback and comments from both the teacher and their peers. To sum up, the integration of emerging technologies including microcontrollers, single board computing and machine learning education tools could empower students to develop AI literacy, cultivate positive attitudes towards AI, foster a gender-inclusive learning environment, and actively engage in collaborative AI exploration.

    Taking into account the perspectives of diversity, equity, and inclusion (DEI), this project primarily focuses on reducing the unequal access to AI learning opportunities among students. It aims to specifically tackle gender disparities by providing tangible refernets and cutting-edge technologies with examples and references that foster a concrete understanding of AI concepts and mechanisms. This project not only aims to enhance the understanding of AI concepts and mechanisms among all students, thereby promoting AI literacy for everyone, but it also places particular emphasis on engaging underrepresented groups such as female or non-binary students in STEM and science subjects. By providing multifaceted ways of emerging technologies enabling students to realize their ideas and thoughts with collaborative learning and real-life solving problems, an inclusive learning environment was designed and implemented to address afformentioned disparities.