Final Statement: AI technologies can be successfully integrated into K-12 mathematics and language arts subjects to enhance student skills and teacher practices when designed through participatory collaboration rather than automation.

Objective: The main goal of this study was to investigate how AI technologies and methods can be integrated into existing K-12 school subjects (mathematics and first language) to support teachers and students in acting skillfully around AI, rather than automating educational tasks that might deskill classroom activities. The researchers aimed to explore how AI can be a driver for skill development within existing subjects through participatory design approaches.

Methods: The researchers conducted a participatory design case study involving collaboration with three schools over several months. The study included two cases: Case 1 involved high school first language teachers (3 teachers) working with AI tools for natural language processing and machine learning, and Case 2 involved elementary school mathematics teachers (6 teachers across two schools) working with machine learning concepts using micro:bit devices and ml-machine.org. The methodology included:

• Collaborative workshops between researchers and teachers to co-design educational tools and activities
• Development of lesson plans integrating AI tools (The Literature Machine, The Engine Room, and ml-machine.org)
• Classroom observations of 23 lessons (90 minutes each) taught by 10 teachers
• Semi-structured interviews with teachers
• Analysis of lesson plans and teaching materials
• Deductive analysis focusing on challenges of integrating AI and how AI interplayed with subject knowledge

Key Findings: The study identified two primary ways AI interplayed with school subjects:

1. Subject knowledge and skills to engage with AI:
   • AI as an object for interrogation: Students used subject knowledge to analyze and discuss AI technologies
   • AI methodologies compared with subject methodologies: Teachers facilitated discussions about quantitative vs. qualitative methods in language analysis
2. AI technologies and methods to become more skillful in the subject:
   • AI to engage with subject content in new ways: Students became highly engaged in text analysis and sentiment annotation
   • AI to expand the scope of subjects: Mathematics teachers introduced concepts like 'data', 'model', and 'algorithm' as mathematical concepts
   • AI as a metaphor to reflect on the subject: AI methods prompted students to reflect on their own reasoning and interpretation processes

The research demonstrated that teachers successfully adapted AI tools to fit their subjects rather than adapting their teaching to the technology. Students showed increased engagement, particularly in text analysis activities, and developed new ways of working with subject content.

Implications: This research contributes significantly to the field of AI in education by providing an alternative to the current trend of automating educational tasks. The findings suggest that:

• Participatory design approaches can effectively integrate AI into existing curricula without deskilling teachers and students
• Different AI tools should be designed for different subjects to support subject-specific skills and knowledge
• AI integration can enhance rather than replace traditional teaching methods
• The approach supports both AI literacy and subject-specific learning simultaneously

The study demonstrates how AI can be positioned as a tool for educational empowerment rather than replacement, aligning with participatory design principles that view humans as "skillful and resourceful in the development of their future practice."

Limitations: The study acknowledges several limitations:

• Time constraints in classroom activities, with teachers reporting insufficient time to make all desired connections to subject content
• Teachers' limited familiarity with AI tools initially required significant preparation time
• Some activities used AI tools without necessarily requiring understanding of AI concepts
• The study involved a relatively small number of teachers and schools
• Some missed opportunities for deeper subject integration due to teachers' limited AI knowledge
• The participatory design approach required extensive collaboration time and resources

Future Directions: The researchers suggest several areas for future investigation:

• Exploring long-term impacts of AI integration in diverse educational settings
• Developing more subject-specific AI tools that support different aspects of various disciplines
• Investigating how to better prepare teachers for AI integration through professional development
• Studying the scalability of participatory design approaches for AI education
• Examining how different interaction techniques and media can support diverse learners in exploring AI concepts
• Researching sustainable implementation strategies for AI integration in schools

Title and Authors: "From Automation to Integration: Designing Opportunities for Students and Teachers to Act Skillfully Around AI in Existing K-12 Subjects" by Karl-Emil Kjær Bilstrup, Luke Connelly, Line Have Musaeus, Magnus Høholt Kaspersen, and Marianne Graves Petersen.

Published on: June 23-26, 2025

Published by: Interaction Design and Children (IDC '25), ACM Conference Proceedings