This paper examines how artificial intelligence could be used to reshape the design of schooling and expand students’ access to high-quality learning experiences for K-12 students in California. It begins from a simple premise: technology matters most when it expands access to the learning experiences that shape long-term outcomes. Rather than evaluating individual AI tools, the analysis begins by establishing the knowledge, skills, and dispositions that matter most for students’ long-term success. It then synthesizes research on the learning experiences shown to cultivate these capacities, draws on qualitative evidence with students, educators, and caregivers to identify five persistent institutional barriers to access, and examines how AI can be deployed as institutional infrastructure to address those barriers and reshape the conditions under which learning occurs. Decades of research show that students’ long-term success in school, work, and civic life depends on a core set of capacities. When schools help students develop deep understanding, strong reasoning, effective communication, self-direction, and the ability to build relationships, they influence how students interpret information, collaborate with others, adapt to change, and continue learning across their lives.
Core Capacities Associated with Long-Term Outcomes
- Academic knowledge and skills
- Higher-order thinking skills
- Social skills
- Metacognition, self-regulation, and adaptability
- Autonomy skills
- Motivation
- Interest and curiosity
- Belonging and interpersonal connection
- Self-efficacy, mindset, and self-concept
- Management of context-specific anxiety, boredom, and frustration
A substantial body of research identifies educational experiences that reliably cultivate these capacities. Across grade levels and contexts, five learning experiences consistently emerge as foundational. When these experiences are present, academic growth strengthens alongside gains in motivation, agency, collaboration, and resilience.
Learning Experiences That Cultivate These Capacities
- Targeted direct instruction responsive to individual learning trajectories
- Real-world learning connected to authentic problems and communities
- Autonomy-supportive environments that structure meaningful choice and student agency
- Enriching discussions that deepen reasoning and perspective-taking
- Caring and supportive relationships characterized by sustained adult mentorship
Yet access to these experiences remains uneven. The constraint is rarely uncertainty about effective practice. Instead, longstanding institutional arrangements shape what is operationally feasible and how opportunities are distributed. Five sets of barriers commonly emerge.
Institutional Barriers Shaping Access
- Institutional structures that determine how time, staffing, and authority are organized
- Assessment and evaluation systems that shape what becomes visible and valued
- Professional capacity constraints that influence how instructional expertise develops
- Curricular and instructional material constraints that affect feasibility and preparation burden
- Communication and equity barriers that shape how opportunities are distributed across students and communities
This paper examines how AI could be designed to address the five persistent institutional barriers that have historically prevented expanded access to meaningful learning experiences. First, within institutional structures, AI can function as a coordination layer that integrates learning data with staffing, space, and scheduling constraints to enable more flexible grouping, targeted support, and time for interdisciplinary, community-connected learning. Second, within assessment systems, AI can support continuous analysis of student work—surfacing patterns in problem-solving, collaboration, and revision—so that academic mastery and competencies such as creativity and metacognition become visible in real time rather than only at the end of a unit. Third, to address professional capacity constraints, AI-enabled simulation and feedback tools can expand opportunities for rehearsal, reflection, and role-specific guidance across educator preparation programs and ongoing professional learning. Fourth, in response to curricular and instructional material constraints, AI systems could align standards-based content to authentic challenges while differentiating for varied readiness levels without imposing unsustainable planning burdens. Finally, to mitigate communication and equity barriers, translation tools, participation analytics, and mentorship-matching platforms can broaden access to relational and experiential opportunities that might otherwise depend on uneven networks or information flows.
Across these domains, the most consequential applications of AI function as integrative systems. They coordinate data, logistics, and professional expertise across the school in service of clearly defined developmental goals. When aligned with research on student capacities and the learning experiences that cultivate them, these systems expand the range of school designs that are operationally feasible at scale. The transformative potential of AI lies in functioning as institutional infrastructure, integrating data, logistics, and expertise across systems that have long operated independently.
AI’s Institutional Levers
- Coordinating time, grouping, and staffing
- Embedding continuous assessment
- Expanding professional rehearsal and feedback
- Connecting standards to authentic inquiry
- Broadening access through communication and matching systems
These applications exist at different stages of development. Near-term tools can be evaluated and deployed within existing infrastructure; the more consequential investment is in systems designed not merely to operate within the current institutional model, but to expand access to the learning experiences that matter most.
The implications extend beyond operational efficiency. When scheduling structures, assessment frameworks, professional learning pathways, curricular systems, and communication networks become more adaptive, students encounter schools organized around responsiveness rather than rigidity. Targeted instruction becomes more feasible. Inquiry has time to unfold. Student voice carries weight in discussion. Mentorship can be sustained rather than fragmented. As schools become more flexible, the experiences that cultivate deep understanding, agency, collaboration, and belonging become more consistently available.
Realizing this potential equitably requires deliberate policy design. Schools serving students with the greatest needs are often least positioned to implement AI-integrated models, and without intentional resource allocation, restructuring risks reproducing existing inequities at scale.
