Co-designing AI curriculum with middle school teachers resulted in effective teaching strategies and improved student engagement across diverse classrooms.

Objective: Study how to develop and implement an effective AI education curriculum for middle school students through collaboration between computer scientists, learning scientists, and middle school teachers.

Methods:

• 3-year study with 8 teachers implementing co-designed curriculum across diverse schools
• 759 students participated in total
• Data collection through classroom observations, teacher interviews, student focus groups
• Analysis of curriculum materials, teaching strategies, and assessment methods
• Used revised Bloom's Taxonomy to evaluate learning objectives and assessments

Key Findings:

• Curriculum structure effectiveness:
  • 52% of content focused on relatable examples
  • Teachers successfully adapted materials for student engagement
  • Activities combining creativity and hands-on work proved most effective
  • Students showed high engagement with ethics discussions
  • Interactive and group-based learning activities were particularly successful
• Assessment strategies:
  • Multiple assessment types aligned with Bloom's Taxonomy levels
  • 30% knowledge-based assessments
  • 26% comprehension assessments
  • 16% creation-focused assessments
  • Unit-long projects effectively reinforced learning
  • Teachers adapted assessments based on student needs
• Co-design benefits:
  • Teachers felt empowered and confident teaching AI concepts
  • Collaboration fostered camaraderie among educators
  • Teachers effectively customized materials for diverse student needs
  • Professional development improved through direct interaction with AI experts

Implications:

• Demonstrates effective framework for teaching AI in middle schools
• Provides model for teacher professional development in AI education
• Shows importance of adaptable curriculum materials
• Highlights value of connecting AI concepts to student experiences
• Establishes assessment strategies for AI literacy

Limitations:

• Study focused only on Georgia schools
• Not all teachers completed full curriculum implementation
• Limited scalability of co-design process
• Potential researcher and teacher bias in materials
• Need for more systematic incorporation of student perspectives

Future Directions:

• Develop online community-driven professional development platform
• Create more creation and evaluation-focused assessments
• Expand study to different geographic regions
• Improve curriculum pacing and implementation support
• Research long-term impact on student AI literacy and career interests
• Design resources for teachers with varied time/resource constraints

Title and Authors: "From Lecture Hall to Homeroom: Co-Designing an AI Elective with Middle School CS Teachers" by William Gelder, Xueru Yu, David Touretzky, Christina Gardner-McCune, and Judith Uchidiuno

Published On: Accepted December 3, 2024

Published By: International Journal of Artificial Intelligence in Education