A Year in K-12 AI Education
Touretzky et. al
https://doi.org/10.1609/aimag.v40i4.5289
In K-12 education, the rise of Artificial Intelligence is crucial as kids are growing up surrounded by it. This disruptive tech has global impacts, prompting efforts in the USA, China, and other countries to integrate AI education in schools. They're creating new curricula and online resources, but the hitch is that most K-12 teachers lack a computer science or AI background.
Challenges include the need for national guidelines in K-12 AI education. Still, a new community is forming, involving AI researchers, educators, and curriculum developers. The AI4K12 Working Group is tracking the progress of AI education.
Over 560 K-12 educators have done an AI course, boosting the US teacher community. Curriculum development is happening worldwide, from China to Finland. The goal is developmentally suitable tools for K-12 AI education, using textbooks, interactive media, and coding platforms.
Teaching AI ethics is gaining importance, focusing on machine learning and bias. Anyone can chip in – join the AI4K12 mailing list, volunteer at local schools, or team up with AI researchers. Terms like AI4K12 Initiative, Cognimates, and Machine Learning for Kids are vital, encapsulating the push for K-12 AI education.
The 3 Concepts
Virtual Reality Data Exploration; Idea #3
Tool/Concept: VR Data Visualization
Description:
-Develop a VR application that allows students to explore datasets in a virtual environment. This can be a fun way to introduce data science concepts. Use VR to create immersive data visualizations, and allow students to interact with the data points. Explain how data is represented in a 3D space, and discuss the importance of understanding patterns and trends in data.
Side Note:
Use features like zooming, rotating, and selecting data points to encourage exploration and understanding.
Explain the significance of different patterns and trends observed during exploration.
Connect the visualizations to real world examples, helping students understand how data impacts decision making.
_________________________________________________________________________________________________________________________________________________
Visual Recognition Game using Apple Vision Pro; Idea #1 & 2
Tool/Concept: Image Recognition Game
Description: Leverage the capabilities of Apple Vision Pro to create a game that involves visual recognition. K-12 students can use their devices to explore their surroundings, and the app will identify objects using image recognition. Introduce machine learning by explaining how the model was trained to recognize various objects. Include a simple leaderboard to add a competitive and engaging element to the game.
_________________________________________________________________________________________________________________________________________________
ML learning for K-12; Idea #1 & 3
Objective:
To introduce students to the concept of computer vision and how machine learning algorithms can identify and categorize objects in images.
Description:
Image Presentation:
Show students a set of pictures featuring objects, terrains, or various scenes.
Each picture in the set comes with a clear description beneath it, helping students understand what is depicted.
Present a new image that is similar to those in the set but not part of it.
Provide multiple-choice options for students to select what the image represents.
If the student selects the correct answer, explain that their decision mirrors how machine learning, specifically computer vision, works. The algorithm recognizes patterns in the data to make accurate predictions.
If the student chooses the wrong answer, show them additional images and repeat the process to reinforce the learning.
DoodleIt: A Novel Tool and Approach for Teaching How CNNs Perform Image Recognition
Mahipal et. al
The DoodleIt application and its integration into an afterschool curriculum for middle school students present a compelling and accessible approach to teaching Artificial Intelligence. The interactive web tool's use of open-source sketch libraries and CNNs for image recognition introduces students to the world of AI in an engaging manner. Its functionality, which allows easy access without installation, demystifies the black box of neural networks, making it a valuable educational resource.
The research underlines the importance of integrating AI education into K-12 curricula, emphasizing the lack of exposure to machine learning processes in children. The AI4K12 Big Ideas, focusing on computer perception, model reasoning, and societal impacts, provide a well-rounded framework for understanding AI's role.
DoodleIt's implementation with CNNs and training using Google's Quick, Draw! dataset highlights practical aspects of AI learning. There's a need for more tools tailored to K-12 students, as addressed by DoodleIt, becomes clear in the study. The research also dives into the challenges faced by students and the subsequent improvements with coaching, demonstrating the adaptability of the curriculum.
A is for Artificial Intelligence
Williams et. al
The research paper provides a comprehensive overview of a study involving the development and implementation of the PopBots early childhood artificial intelligence platform for preschool children. The assessment of student learning outcomes focused on children's understanding of AI concepts through hands-on activities. The results indicated a median score of 70% on the cumulative assessment, with knowledge-based systems being better understood.
The study effectively explored the impact of AI activities on children's perceptions of robots, revealing that younger children tended to view robots as smarter toys, while older children saw them as less intelligent people. The research emphasized the importance of early AI education to empower children in a rapidly changing, AI-powered world. It recognized that children already interact with AI devices daily and lack understanding, leading to faulty assumptions about their capabilities.
The hands-on activities using the PopBots platform, including machine learning and reasoning algorithms, were designed to teach children about various AI concepts. The assessments, including Knowledge-Based Systems, Supervised Machine Learning, and Generative AI, measured their understanding. The research considered the impact of developmental factors, such as Theory of Mind skills and age, on children's learning about AI.