In Spring 2025, FullScale (formerly The Learning Accelerator) identified a critical need in the sector: leaders need support navigating and making sense of AI. Leveraging learnings from two of our major projects, the Exponential Learning Initiative and the School Teams AI Collaborative, we surfaced a driving question: What would it take for AI to transform K-12 learning at the systems level, not just automate what we already do?
For AI to be transformational, we have to push past the typical conversations about tool features, efficiency gains, and compliance risks. While these issues are critically important, if the conversation stalls there, we fail to explore how AI could impact the instructional core—the interactions between students, teachers, and content. Without research and policy focus in this area, this watershed technology risks becoming another add-on that makes schools more efficient without fundamentally changing what students actually experience or learn.
With this framing in mind, we took our driving question to trusted partners, state and local education leaders, researchers, and funders to better understand the vision for transformational teaching and learning with AI. Through focus groups, design sprints, and ongoing conversations, we tested ideas, refined insights, and began to shape a new set of resources based on what we learned.
Through this process, three clear opportunities emerged where researchers and policymakers can help move the field forward with AI.
1. Center Research on the Instructional Core
Richard Elmore defined the instructional core as the interactions between students, teachers, and content. In our School Teams AI Collaborative and with some of the networks participating in our Exponential Learning Initiative’s Accelerating Adoption Network, we supported educators and teams to make durable changes to the core of instruction by co-designing plans to track impact through student engagement, work products, and assessment data. This stood in contrast to novelty-driven uses of AI, such as auto-generating worksheets or automating grading, which may save time but do little to change how students engage with content or demonstrate their learning.
With this support, teams tested and iterated on AI-integrated practices and gathered evidence about what worked, what didn’t, and under which conditions. Teachers reported richer feedback cycles and deeper engagement, and early data pointed to gains in student performance. For researchers and policymakers, this reinforces the importance of grounding inquiry and investment in the instructional core, ensuring that AI studies and strategies demonstrate changes in learning and not just tool use.
2. Understand the Conditions that Support Innovation
Through the Exponential Learning Initiative, we partnered with leaders to map the contexts that shape whether AI efforts thrive or falter. Factors ranging from infrastructure (such as devices, bandwidth, and secure data systems) and scheduling to professional learning and leadership vision often determined whether AI-enabled practices sustained and scaled or failed to gain traction.
To support this work, we developed our Starting Smart with AI guides for both state and local leaders. These resources outline foundational conditions—such as a clear instructional vision, strong data privacy protections, educator support, and bias mitigation—that can be addressed as early steps. By grounding work in these essentials, state and local education agencies can build the structural conditions needed for transformative teaching and learning with AI. At the same time, our work highlighted the need for research that connects instructional outcomes to the enabling conditions that make them possible, so policymakers and practitioners can better understand how to replicate success across contexts.
3. Strengthen Coherence Across the Field
Across the country, educators are already adapting their practice and seeing glimpses of what AI can make possible. However, a critical factor remains missing: the infrastructure and coherence to connect these isolated efforts, capture lessons in real time, share them across contexts, and turn them into strategies that others can use.
This insight informed the development of From Urgent to Future: Charting A Course for AI in K-12, our resource suite designed to build coherence across the sector. By aligning research, practice, and policy around a shared evidence base that is timely, relevant, and shaped by practitioners, we can accelerate progress while avoiding fragmentation. For researchers and policymakers, this underscores the need to invest in infrastructure that connects learning across systems, helping the sector build coherence rather than a patchwork of isolated experiments.
Resources to Support Turning Opportunity into Action
Across this work, we have seen that progress with AI depends on much more than isolated classroom pilots. It requires research and policy that help systems focus on instruction, understand enabling conditions, and stay connected across contexts. To support this, in addition to the resources highlighted above, we developed a set of coordinated action recommendations, designed to support leaders across and throughout the sector in making good on these opportunities.
We consider these resources as starting points for building the shared evidence base and policy coherence that the field needs. Researchers, policymakers, and practitioners each have a role to play in making good on the opportunities surfaced here and in ensuring AI is used to strengthen, not sideline, powerful teaching and learning.
