Unpacking Approaches to Learning and Teaching Machine Learning in K-12 Education: Transparency, Ethics, and Design Activities
Project Overview
The document explores the integration of generative AI and machine learning (ML) into K-12 education, identifying three key approaches to teaching ML: a data-driven approach that focuses on dataset creation and its influence on models, a learning algorithm-driven approach that emphasizes understanding algorithms, and an integrative approach combining both methods. It underscores the necessity of incorporating ethics and algorithmic justice into ML curricula to ensure that students not only learn the technical aspects but also understand the societal implications of these technologies. The document also addresses significant challenges, including the inherent opacity of ML algorithms, and calls for the development of accessible educational tools designed for younger learners to enhance their comprehension of complex concepts. Overall, the findings suggest that while there are promising applications of generative AI in educational settings, careful consideration of ethical frameworks and pedagogical strategies is essential to foster a well-rounded understanding of AI among students.
Key Applications
Interactive Machine Learning and Neural Network Learning
Context: K-12 education, targeting students from early childhood to high school, including middle and high school students with varying levels of coding experience.
Implementation: Students engage in hands-on activities where they create and label datasets, train machine learning models, and conduct coding exercises and projects. These activities often focus on real-world applications such as environmental issues or social studies, and utilize existing datasets to enhance understanding of algorithms.
Outcomes: Improved conceptual understanding of machine learning principles and algorithms, increased engagement through personally relevant projects, and a deeper understanding of both data creation and algorithm application.
Challenges: Limited integration of ethics and critical discussions around algorithmic justice, lack of opportunities for students to design their own datasets, and a need for more comprehensive tools that support both dataset creation and algorithm exploration.
Implementation Barriers
Technical Barrier
Opacity of machine learning algorithms makes it difficult for students to understand how models work.
Proposed Solutions: Develop tools that visualize the inner workings of algorithms and make learning resources more transparent.
Pedagogical Barrier
Limited integration of ethics and algorithmic justice into current ML curricula.
Proposed Solutions: Incorporate ethics discussions into technical learning activities and encourage students to critically evaluate their own projects.
Curricular Barrier
Many existing ML educational programs focus on data or algorithms separately without integrating both effectively.
Proposed Solutions: Create curricula that combine both approaches, exploring how data and algorithms interact in machine learning, and include ethical considerations.
Project Team
Luis Morales-Navarro
Researcher
Yasmin B. Kafai
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Luis Morales-Navarro, Yasmin B. Kafai
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