As I reflect on the past year, I am grateful for our partners who are pioneering new ways to improve teaching and learning. Their work inspires us to develop the research-driven building blocks that help scale quality teaching and learning practices via AI-powered edtech tools. Together, we can make tools and resources that are more impactful for teachers and students.
This year, we launched an early release of Knowledge Graph and Evaluators and partnered with Anthropic to connect Knowledge Graph to Claude. Last month, we announced updates to Knowledge Graph, including the integration of math learning standards from Washington, D.C., as well as alignment between math learning components (smaller skills and concepts) and academic standards for Louisiana, Montana, Pennsylvania, and Washington, D.C. We will continue enhancing these tools as we build toward general availability in 2026.
As excited as I am about our progress this year, I also recognize that our work today builds on a strong foundation of research to codify and advance the science of learning and development; of partnerships to put this research into practice; and of collaborative initiatives to design tools with educators and researchers — including efforts to thoughtfully explore the potential of AI tools in schools.
Since our earliest grants a decade ago, we have worked to translate learning science into accessible instructional practices and classroom tools. For example, in 2019, we partnered with the Gates Foundation and NewSchools Venture Fund to launch EF+Math, a $50 million initiative aimed at improving math outcomes for students in grades 3-8. The program supported teams of educators, researchers, and developers working to strengthen students’ executive function skills — like self-control and flexible thinking — which research shows can accelerate math learning. That initial effort was part of the broader Advanced Education Research & Development Fund (AERDF), which applies research and development methods that center educators and communities in co-designing solutions to challenges in education. Today, school districts and education systems nationwide are experiencing the impact of AERDF’s tools and strategies.
As machine learning and artificial intelligence tools became more accessible, we partnered with organizations exploring the potential of these new technologies for education. We supported Quill to help more than one million students strengthen their writing skills through AI-powered feedback grounded in learning science. We also backed a partnership between the Learning Agency Lab and Georgia State University to launch The Feedback Prize. The project leverages AI to develop free tools that help teachers supplement their feedback to students struggling with writing by providing more frequent and automated feedback.
Over the past couple of years, we have expanded on these initial efforts by supporting the development of openly licensed datasets that improve how AI performs in education, funding projects like the AI x Coherence Academy to equip education leaders to thoughtfully integrate generative AI in improvement efforts, and backing initiatives to empower educators as co-creators of future technologies.
Today, we are working with longstanding partners to incorporate their research into our developer tools. For example, we are supporting Quill to create a high-quality database of feedback on student writing that will strengthen evaluations of AI output. We are also leveraging research on foundational literacy skills developed by AERDF’s Reading Reimagined program, via Magpie Literacy, to improve Knowledge Graph. Similarly, we are working with Digital Promise to leverage the Learner Variability Navigator, a resource we funded that helps educators find research-based instructional strategies to meet the unique needs of students — in our infrastructure tools.
In the coming years, we will expand our partnerships with curriculum providers, researchers, and educational organizations to develop open public infrastructure to support AI tools that scale proven teaching and learning practices. Playlab exemplifies how this infrastructure is being used. Using Knowledge Graph, Playlab connects its platform directly to learning science, curriculum, and academic standards — giving teachers the agency to shape how AI shows up in their classrooms and helping them plan lessons, unpack standards, and focus on students’ growth.
Over the past decade, we’ve learned a lot about the challenges and opportunities in our education system (for more about our work in the early years, read the report on our first eight years). This past year has shown us what is possible as we turn our focus to the instructional infrastructure our field needs. We are so inspired by the response and product integrations we’ve already seen from edtech partners bringing better capabilities to educators built with our tooling. As we look ahead, we continue to be guided by the belief that every learner matters and that the right tools can help teachers transform students’ lives for the better.
Sandra Liu Huang
President, Learning Commons