The Measure of Learning: How Independent Schools are Navigating AI Through Policy, Pedagogy, and Innovation

Abstract

Artificial intelligence has entered school life more quickly than institutions have developed shared expectations for its use. Independent schools now face connected questions about what students should learn, which intellectual work they must undertake themselves, how teachers should evaluate their understanding, and who is responsible for decisions about emerging tools. Those questions reach across classrooms, academic leadership, student life, and institutional operations. The Measure of Learning examines these responsibilities through published research and CONSYLIUM’s 2025–26 advisory partnerships with Cheshire Academy, Kimball Union Academy, and Woodberry Forest School. The engagements comprised 45 one-hour sessions involving faculty, school leaders, and students. Their different institutional priorities illuminate the decisions schools face as they respond to AI while maintaining their educational commitments. Across the partnerships, faculty judgment shaped the central instructional questions. Humanities teachers examined authorship, interpretation, and the development of an argument. STEM teachers emphasized verification, explanation, and testing AI-generated answers against evidence. Across disciplines, faculty considered how to make student reasoning visible and assess understanding when AI contributes to a finished product. Sustained, discipline-specific professional development gave teachers room to examine these questions in the context of their own practice. Student innovation extended the inquiry into practical work. Through a model focused on service to the school community, students developed early projects that required them to identify needs, consult stakeholders, consider privacy and reliability, and account for their design decisions. Exchanges with peers at other schools introduced additional perspectives and possibilities. These experiences showed how students can collaborate on campus problems and contribute to discussions of AI policy. The policy work emphasized the importance of grounding institutional expectations in a school’s educational philosophy. Commitments to learning, character, and relationships informed different approaches to AI across the partner communities. An institutional stance can establish principles to guide future decisions, while policy translates those principles into expectations that faculty, students, and families can understand and apply. Governance connects those expectations to institutional responsibility. Schools must determine who approves tools, how faculty and staff AI use is governed, what safeguards apply to sensitive information and consequential decisions, and how guidance will be maintained as technology changes. Centralized and distributed models offer different ways to organize oversight, but both require clear accountability and the involvement of educators. Social AI and synthetic media also bring questions of student well-being and community trust into this work. Twelve recommendations translate these findings into practical actions for independent schools across institutional leadership, instructional design, student participation, and community well-being. Looking Ahead introduces CONSYLIUM’s research partnership with the Child-Centered AI Lab at Harvard Graduate School of Education, established to support rigorous inquiry into how AI shapes young people’s learning, development, and well-being and to generate evidence that can inform decisions within participating school communities.