Teaching, Learning, and AI in the K-8 Classroom

Generative artificial intelligence, or gen AI, is a rapidly advancing set of technologies making its mark in many fields and industries, including education. The launch of ChatGPT in November 2022 was a breakthrough moment with global impact. "At a faculty meeting the spring after ChatGPT was released, Mark Day teachers were already asking themselves what gen AI would mean for teaching and learning," says Bonnie Nishihara, Assistant Head and Director of Educational Design & Innovation at Mark Day School. "We have and will continue to ask that question over and over again as gen AI continues to evolve. The impact of AI will be deep, far-reaching, and lasting; it's not just a passing fad."

Almost immediately, we discovered that the big ideas and guiding principles that have served us through decades of technology integration at Mark Day—through laptop and iPad adoption, through media production and design and tinkering program development—remain deeply relevant in the age of AI. For example, we believe in using technology as a creative tool (or, as we like to say, as “fingerpaint”). In March 2023, Design, Tinkering, and Technology Integrator Tatian Greenleaf launched the first of many elective courses for 7th and 8th graders focused on using gen AI to produce creative works. In the class, students not only used gen AI tools to create text, generate images, and write code, but also examined how gen AI is trained and the ethical, legal, and environmental issues it raises.

We also believe that technology is a tool that can support students in achieving educational goals. Sometimes it's the right tool for a task; sometimes it isn't. Either way, technology is always just a tool, and we are intentional about how we use it. (To get this point across, I often say, “A tool is a tool is a tool is a tool.”) For students in K-8, gen AI is rarely the right tool because students at this age are still mastering foundational skills such as reading, writing, research, and mathematical and critical thinking. Author, professor, and researcher Ethan Mollick describes the “Jagged Frontier” of AI: AI's performance on different types of tasks is remarkably variable, or jagged. ChatGPT can score in the 90th percentile on the bar exam but gets basic algebra problems wrong a third of the time. With that variability in mind, Mollick argues that gen AI tools need a skilled "human in the loop"—a user who can assess the reliability of output, apply human judgment, and make informed decisions throughout the process. "It is exactly this—training the 'human in the loop'—that Mark Day is highly focused on and perfectly positioned to do," says Head of School Joe Harvey.

In the fall of 2023, Mark Day teachers in grades 5-8 participated in a series of workshops designed to deepen their understanding of gen AI. Through hands-on use of gen AI tools, teachers began learning firsthand about the “Jagged Frontier” of gen AI’s capabilities. It was a time when many schools were narrowly focused on banning access to gen AI. At Mark Day, banning access is what we call a “guardrail” solution. Guardrails are important; Mark Day’s content filter blocks many gen AI tools by default, too. Mark Day teachers, however, were also asking the deeper question of how teaching and learning in the classroom may need to shift to ensure we are also acting as “guides” for students. For example, we teach writing as a multi-step process to guide students to develop essential, durable skills. We are exploring how we strengthen that approach in a manner that is more resistant to overreliance on AI than a guardrail solution alone. Further, what role, if any, should gen AI tools play in skill development?

In August, Mark Day hosted a four-day summer institute titled "Teaching, Learning, and AI in the K-8 Classroom," attended by teachers from Mark Day and other independent schools. Led primarily by GOA (Global Online Academy), the Institute combined keynote speakers, hands-on workshops, and design time during which teachers developed plans for the 2024-2025 school year. Participants learned how to use gen AI to support their own work—particularly in differentiation and assessment—while wrestling with our core "a tool is a tool" questions about how they might thoughtfully incorporate gen AI into their teaching in age-appropriate ways to meet educational goals.

Workshops included:

Intro to AI in Education

This interactive, level-setting session supported educators in better understanding AI’s implications on learning and classroom practice. Participants experienced foundational understanding, practiced with fundamental AI tools and processes, and had exploration/design time. Participants learned concrete strategies for leveraging AI tools that support their roles as educators

Assess for Learning with AI

In this session participants prioritized what it means to assess for learning and how AI can support student-centered assessment. This workshop offered level-setting experiences, design sprints, support, and coaching in designing for student-centered assessment practices with the support of AI. Participants got to learn and apply strategies for small shifts in approaches to feedback, reflection, demonstrations of learning, and formative assessment that integrate AI and that prioritize student agency. Participants left with replicable approaches to redesigning assessment for learning.

AI and Differentiated Learning

Learning can be more meaningful and empowering when learners engage in experiences that are responsive to their individual needs and relevant to their interests. In this session, participants discovered how to use AI platforms to differentiate learning, focusing on designing for more agency, greater depth, and increased relevance. After collaborative exploration and practical application, participants left more equipped to develop responsive and relevant learning experiences for each student with the support of AI.

Student Use of AI - What to Know, Think, Do

Student-facing AI use holds both promise and limitations. In this introductory session participants explored must-knows, use considerations, and helpful strategies for leveraging AI in partnership with students (and aligned to their age access levels).

Design a Feedback Ecosystem

Assessments that are rooted in robust cycles of feedback have the single most significant impact on student learning. To ensure that educators can maximize the effectiveness of feedback while managing feedback workload, they need to employ the use of a feedback ecosystem. This session explored the what, why, and how of providing students with ongoing, timely, and actionable feedback as well as AI tools to integrate for more efficiency and effectiveness. Participants left with a plan for launching a learning experience with a feedback ecosystem in place.