Target Audience
- Educators, researchers, and professionals interested in AI in education, EdTech developments, and the intersection of generative AI and learning science.
- Developers and organizations building or adopting AI-powered educational services, particularly those concerned with educational quality and learning outcomes.
- Policymakers, nonprofit organizations, and researchers interested in open educational infrastructure, public-private collaboration, and the governance of educational technology.
Summary
- Learning Commons is an initiative that reorganizes approximately a decade of educational research and technology investments by the Chan Zuckerberg Initiative (CZI) into shared AI infrastructure for education.
- The name Learning Commons was formally introduced in September 2025, and an open AI development platform for education providers and researchers was launched on September 22, 2026.
- Through Knowledge Graph, Evaluators, Agent Skills, and Curriculum Sync, Learning Commons provides structured educational knowledge, quality assessment tools for AI-generated content, reusable instructional procedures, and connections between curricula and learning management systems.
- The initiative collaborates with OpenAI, Anthropic, education technology companies, research institutions, and school districts. However, it faces challenges involving evidence of learning outcomes, governance of educational standards, implementation costs, international expansion, and long-term sustainability.
- As of November 2026, Learning Commons has established and publicly launched its shared infrastructure and is expanding its partnerships. Demonstrating measurable educational impact and earning sustained public trust remain important next steps.
1. Introduction: What Is Learning Commons?
Learning Commons is a nonprofit initiative supported by the Chan Zuckerberg Initiative (CZI) that aims to build shared infrastructure connecting educational research with generative artificial intelligence.
CZI was established in 2015 by Mark Zuckerberg and Priscilla Chan. Since its founding, the organization has pursued long-term investments in areas including education and science.
Learning Commons represents an effort to reorganize approximately a decade of CZI's educational research and technology investments into shared infrastructure designed for the era of generative AI.
The name Learning Commons was formally introduced in September 2025. On September 22, 2026, the initiative launched an open AI development platform accessible to education providers and researchers. [1][2]
The official Learning Commons website is available at:
Led by President Sandra Liu Huang, Learning Commons provides educational data, evaluation technologies, and AI-oriented instructional design resources as nonprofit infrastructure intended to serve the public interest.
As of September 22, 2026, the initiative reported more than 70 partners contributing data or participating in technical collaborations. [2][3]
Five Key Points About Learning Commons
Five aspects are particularly important for understanding Learning Commons.
-
The focus of CZI's education strategy has shifted. Rather than primarily supporting schools, teachers, and individual educational programs directly, the initiative increasingly focuses on infrastructure that multiple organizations developing educational AI can use.
-
It connects learning science with AI. Learning Commons transforms educational standards, curricula, prerequisite relationships between learning concepts, student misconceptions, and instructional approaches into data that AI systems can use.
-
It provides infrastructure rather than exclusively delivering end-user AI products. Instead of seeking to be the sole provider of applications for teachers, Learning Commons aims to supply capabilities that can be integrated into multiple educational services and AI platforms.
-
It combines openly available resources with strong technical partnerships. Its collaborators include OpenAI, Anthropic, education technology companies, research institutions, and school districts.
-
Its central challenge is demonstrating that improvements in educational AI quality lead to better learning outcomes. Additional questions concern who determines educational standards, who evaluates instructional quality, and how the influence of a privately funded philanthropic organization over educational infrastructure should be governed.
Although the platform launch in September 2026 represented an important milestone, it did not mean that Learning Commons had already achieved widespread adoption or conclusively demonstrated educational effectiveness in schools.
2. How Did Learning Commons Emerge?
Learning Commons did not suddenly appear as an entirely new initiative in 2025.
It is more accurately understood as the result of several strategic shifts in CZI's education activities.
Its development can be divided into four broad stages. [1][7]
2.1. Stage One (2010–2016): Direct Investment in Education Reform and Educational Opportunity
The history leading to Learning Commons includes Mark Zuckerberg's support for education reform in Newark, New Jersey, beginning in 2010.
Following CZI's establishment in 2015, investments expanded into school reform, educational opportunity, educational research, and educational technology.
During this period, the organization primarily supported efforts addressing challenges directly within schools and educational systems, including specific school improvement programs and initiatives intended to expand educational opportunities.
2.2. Stage Two (Approximately 2016–2022): Personalized Learning and the Application of Learning Science
In the next stage, CZI increased its focus on individual differences among learners, learning science, and social and emotional development.
One representative initiative was its support for educational programs such as Summit Learning.
CZI also supported educational technologies used by schools and teachers.
Its activities during this period addressed a wide range of challenges, including personalized learning, teacher support, social-emotional learning, and support for students across school and home environments.
2.3. Stage Three (2023–2024): A Shift Toward Foundational Technologies for Educational AI
Beginning in 2023, the growing adoption of generative AI made the standardization of educational data and the quality assessment of AI-generated instructional materials increasingly important.
Against this backdrop, CZI's education strategy began shifting toward connecting AI technologies with learning science.
On December 16, 2024, CZI announced private beta versions of Knowledge Graph and Evaluators. [4]
These technologies were designed to transform educational knowledge into structured data that AI systems could use and to evaluate the educational quality of AI-generated content.
This marked an important transition.
Rather than focusing solely on directly supporting schools and teachers using educational AI, CZI began developing shared infrastructure that education technology companies and researchers could use to build their own AI-powered tools.
2.4. Stage Four (2025–2026): Consolidation and Public Launch of Learning Commons
On September 23, 2025, the name Learning Commons was announced, and initial datasets and related resources began to be released. [1]
This clarified a new organizational direction centered on shared resources for educational AI.
During July and August 2026, collaborations with Anthropic and OpenAI in educational AI expanded.
On September 22, 2026, Learning Commons formally launched its developer-facing platform. [2]
In October 2026, the initiative continued publishing information about responsible AI, critical thinking, and related topics, further extending its work on the quality and evaluation of educational AI.
2.5. Timeline of Major Developments
| Period | Event | Significance |
|---|---|---|
| 2015 | CZI was established. | Long-term investments in education, science, and other fields began. |
| 2016–2022 | Investments in learning science, personalized learning, and educational technology. | Research, practical knowledge, and technical capabilities were accumulated. |
| 2023 | Education division restructuring and staff reductions. | CZI began revising its previously broad education strategy. |
| December 16, 2024 | Private beta announcement of Knowledge Graph and Evaluators. | Shared infrastructure for educational AI developers began taking concrete shape. |
| September 23, 2025 | Announcement of Learning Commons and initial datasets. | A new organizational direction centered on common infrastructure became visible. |
| July–August 2026 | Expansion of educational AI collaborations with Anthropic and OpenAI. | Learning science began being integrated into general-purpose generative AI services. |
| September 22, 2026 | Public launch of the developer platform. | The initiative entered a new phase of public access and partnership expansion. |
| October 2026 | Continued work and communication on responsible AI and critical thinking. | Its activities related to educational AI quality and evaluation continued to expand. |
2.6. Understanding the Strategic Shift in CZI's Education Activities
A central point in understanding Learning Commons is CZI's transition from supporting individual programs that address educational problems to building shared technological infrastructure used by multiple educational services.
Previously, CZI worked on a broad set of educational issues, including personalized learning, teacher support, social-emotional learning, and connections between schools and families.
However, staff reductions affecting 48 positions in its education division were reported in 2023. During 2025 and 2026, CZI also continued concentrating its organizational priorities. [7]
Compared with its earlier direct support for educational programs, Learning Commons focuses more specifically on the intersection of AI technology and educational research.
There are at least two ways to interpret this shift.
CZI's explanation
CZI has emphasized that approximately a decade of investment in learning science created a body of knowledge that generative AI can now help deliver to schools on a much broader scale. [1]
An external interpretation
Another perspective is that CZI reduced its support for a broad range of educational opportunity programs and grantmaking activities while concentrating resources on areas aligned with its strengths in software development, AI, and data infrastructure. [7]
These interpretations are not necessarily contradictory.
However, publicly available information does not establish every aspect of the organization's internal decision-making or the motivations behind its strategic changes.
3. How Learning Commons Works and What Makes It Distinctive
The basic concept behind Learning Commons is to make knowledge from educational research and learning science available as shared infrastructure that AI systems and educational technology developers can reuse.
Traditionally, education technology companies and organizations have often needed to independently manage tasks such as aligning their products with educational standards, organizing instructional data, and evaluating the quality of teaching materials.
Learning Commons seeks to provide common resources for these activities, supporting the development and application of educational AI.
3.1. Overview of the Learning Commons Architecture
The overall structure can be summarized as follows.
Educational Research, Learning Science,
Educational Standards, and Curricula
|
| Knowledge about what is
| educationally appropriate
| and effective
v
+--------------------------------------+
| Learning Commons |
| Shared Infrastructure for AI |
| in Education |
| |
| Knowledge Graph |
| Structures educational knowledge |
| |
| Evaluators |
| Assesses AI-generated content |
| |
| Agent Skills |
| Provides reusable instructional |
| procedures for AI agents |
| |
| Curriculum Sync |
| Connects curricula with LMSs |
+--------------------------------------+
|
v
AI and EdTech Companies, Schools,
School Districts, Curriculum Developers
|
| Used by services such as
| ChatGPT, Claude, and
| specialized educational AI
v
Educational Services Used by
Teachers and Students
|
v
Lesson Planning, Content Creation,
Feedback, and Learning Support
This diagram is a conceptual representation based on official Learning Commons materials.
The four products do not necessarily operate as consecutive steps in a single pipeline. They can be combined according to different use cases.
3.2. Knowledge Graph: Structuring Educational Knowledge
Knowledge Graph organizes educational knowledge into structured data that AI systems and developers can access.
Its scope includes educational standards from all 50 U.S. states, learning concepts, prerequisite relationships among learning objectives, and curriculum information.
By September 2026, Learning Commons had introduced resources covering K–12 mathematics, English language arts learning concepts for kindergarten through second grade, and datasets from Eedi addressing more than 8,000 mathematical misconceptions. [2][5]
Potential applications include:
- Generating questions aligned with educational standards.
- Identifying prerequisite knowledge needed to understand a learning concept.
- Recommending learning materials based on students' current levels of understanding.
An important characteristic of Knowledge Graph is that it does not treat educational content merely as a collection of text.
Instead, it structures relationships among learning concepts, including dependencies and instructional sequences.
3.3. Evaluators: Assessing the Quality of AI-Generated Educational Content
Evaluators assesses AI-generated teaching materials and feedback using criteria developed in collaboration with educational experts.
Its evaluation dimensions include:
- Grade-level appropriateness.
- Text complexity.
- Alignment with mathematics learning standards.
- Quality of writing feedback.
- Critical thinking.
Rather than reducing educational quality to a single overall score, the system evaluates distinct aspects of quality. [5]
Potential applications include quality assurance for AI-generated instructional materials, model and prompt improvement, and testing before product releases.
For educational AI, generating grammatically correct and seemingly appropriate content is not enough.
For example, an AI system may generate a passage intended for third-grade students. However, whether that passage actually matches the intended reading level, aligns with instructional objectives, and supports student understanding are separate questions.
Evaluators is designed to assess educational qualities that cannot be adequately captured by general measures of text generation performance alone.
3.4. Agent Skills: Integrating Instructional Practices into AI
Agent Skills provides reusable procedures that allow AI agents to apply knowledge developed by educational experts.
Examples of instructional activities include:
- Creating lesson plans.
- Adapting lessons to different learners.
- Preparing classroom instruction.
- Checking student understanding.
As of September 2026, four Agent Skills developed in collaboration with Anthropic had been released, with additional development also announced. [2][5]
Potential applications include educational AI agents, teaching assistants powered by AI, and tools supporting lesson preparation.
While Knowledge Graph provides structured educational knowledge, Agent Skills addresses how AI systems can use such knowledge to carry out instructional tasks and support teaching processes.
3.5. Curriculum Sync: Connecting Curricula with Learning Management Systems
Curriculum Sync helps schools use their selected teaching materials and curricula within existing learning management systems (LMSs).
It aims to facilitate the use of educational materials in platforms such as Canvas and Google Classroom.
Its primary applications include:
- Making curriculum adoption more efficient.
- Sharing instructional materials through LMS platforms.
- Supporting collaborative development with school districts.
A key feature is its effort to reduce the manual work involved in transferring educational content and make materials easier to use within existing systems. [5]
3.6. Five Design Principles That Distinguish Learning Commons
The design philosophy of Learning Commons can be summarized through five characteristics.
| Characteristic | Difference from Conventional EdTech |
|---|---|
| Shared infrastructure | Provides capabilities that multiple applications can integrate, rather than serving only as a standalone application for particular schools. |
| Emphasis on learning science | Prioritizes educational validity in addition to fluent responses from general-purpose AI. |
| Openness and reuse | Makes datasets and open-source resources available for broader reuse. |
| Emphasis on evaluation | Assesses the educational quality of AI-generated content, not merely the ability to generate it. |
| Collaborative development | Integrates knowledge from researchers, educational institutions, companies, and practicing teachers. |
One of the most important distinctions is the difference between making AI answer educational questions effectively and designing AI systems that genuinely support learning.
Learning Commons aims to address educational qualities that general-purpose AI systems may not adequately capture through three complementary approaches: educational data, quality evaluation, and instructional procedures. [1][5]
4. Current Status and Applications of Learning Commons
With the platform launch in September 2026, Learning Commons moved beyond initial concepts and limited pilots into a phase in which multiple developers could access and use its infrastructure.
4.1. Key Scale Indicators as of September 2026
| Indicator | Reported Figure | Description |
|---|---|---|
| U.S. state educational standards | All 50 states | Standards data covering four major subjects. |
| Dataset downloads | Approximately 20,000 | Reported by EdSurge in September 2026. |
| Partners | More than 70 | Includes data contributors and technical integration partners. |
| Publicly released Agent Skills | Four | Main publicly available set as of September 2026. |
These figures reflect information reported in September 2026 and should not be interpreted as updated cumulative figures for November 4, 2026. [2][3][6]
Importantly, dataset downloads do not directly indicate the number of schools actively using Learning Commons or improvements in student learning outcomes.
4.2. Collaborations with Anthropic and OpenAI
Learning Commons has been working with Anthropic and OpenAI to incorporate educational knowledge into generative AI systems.
Knowledge Graph has been connected with education-oriented services such as Claude for Teachers and ChatGPT to support the creation of teaching materials based on educational standards and learning concepts.
Learning Commons has also collaborated with Anthropic to develop Agent Skills. [2][5]
These relationships demonstrate that Learning Commons is not simply developing its own educational AI application.
Instead, it aims to become shared infrastructure supporting educational applications built on general-purpose AI platforms.
4.3. Canva Education and MagicSchool
Canva uses Knowledge Graph to support the creation of instructional materials aligned with educational standards.
MagicSchool is also collaborating with Learning Commons to incorporate educational standards and structured knowledge into its AI services. [2][5]
These examples illustrate how Learning Commons can support not only general-purpose AI platforms but also companies specializing in educational technology.
4.4. TalkingPoints
TalkingPoints uses educational data from Knowledge Graph to connect classroom learning with support provided at home.
It supports learning activities for families translated into more than 150 languages. [5]
This example demonstrates that applications of Learning Commons can extend beyond instructional content generation and lesson preparation to supporting connections between schools and families.
4.5. Lisle District 202 in Illinois
Lisle District 202 in Illinois used a beta version of Knowledge Graph to prototype a mathematics formative assessment application.
The district explored how prerequisite knowledge and educational standards could help educators better understand students' learning progress. [6]
However, the district reported that additional development support was needed for features required in production, including authentication, dashboards, and hosting.
This example demonstrates both the practical potential of Learning Commons and an important limitation.
Even when shared infrastructure is available at no cost, turning it into a fully operational application for schools still requires technical development and resources.
5. Challenges Facing Learning Commons
Learning Commons offers significant possibilities as shared infrastructure for educational AI.
At the same time, it faces six major challenges: demonstrating educational effectiveness, establishing governance, reducing implementation burdens, addressing limitations of AI-based evaluation, supporting international adoption, and ensuring long-term sustainability.
The following sections distinguish concerns documented by external stakeholders from risks that can be inferred from the initiative's operating model.
5.1. Challenge One: Demonstrating Educational Effectiveness
Learning Commons is designed to improve the quality of educational AI.
However, improvements in the quality of AI-generated teaching materials do not automatically translate into improved student learning outcomes.
For example, even when generated questions align with educational standards, it remains necessary to examine whether they improve student understanding or long-term knowledge retention.
Publicly available information as of 2026 includes usage data, developer feedback, and examples of pilot implementations.
However, sufficient evidence of large-scale independent evaluations demonstrating a causal effect of Learning Commons on academic achievement has not been identified.
Important indicators for future evaluation include:
- Efficiency of instructional content creation.
- Agreement between automated evaluations and expert judgments.
- Changes in teachers' workloads.
- Actual student learning outcomes.
Connecting improvements in technical quality with measurable educational outcomes will therefore be a central challenge.
5.2. Challenge Two: Governance of Educational Knowledge and Evaluation Standards
In its September 22, 2026 reporting, EdSurge described concerns raised by university researchers and education stakeholders about the role of a privately funded philanthropic organization in controlling infrastructure for educational AI. [6]
Several important governance questions emerge:
- Which research findings and teaching materials should be considered trustworthy enough to include?
- Which instructional approaches and theories of learning should be treated as desirable?
- Who makes decisions when educational content or evaluation standards are revised?
- How should competing educational theories and perspectives be represented?
Learning Commons has published principles concerning data governance and responsible AI. [8]
This is an important step.
However, openness alone does not necessarily resolve concerns about the concentration of influence over shared educational infrastructure.
Because shared AI infrastructure may affect a wide range of educational services, transparency and accountability in decisions concerning datasets and evaluation standards are essential.
5.3. Challenge Three: Implementation Burdens for Schools
Even when open educational datasets and APIs are available, schools may struggle to use them without sufficient technical expertise.
The experience of Lisle District 202 illustrates this problem.
Although Knowledge Graph helped the district develop a prototype, additional technical expertise and external assistance were required to implement authentication, management interfaces, and related features. [6]
In other words, providing shared infrastructure free of charge does not eliminate the cost of developing and operating applications that can be used in real classrooms.
Schools with limited financial or technical capacity may therefore need more than access to shared infrastructure.
Implementation support and user-friendly, ready-to-use applications may also be necessary.
5.4. Challenge Four: Limitations of AI-Based Evaluation
Evaluators uses criteria developed with educational experts to assess the quality of AI-generated outputs.
However, some aspects of educational quality do not have a single correct answer.
Automated assessments may not fully reproduce the judgments of experienced educators.
In particular, abilities and characteristics such as the following are difficult to capture solely by analyzing the surface quality of written content:
- Critical thinking.
- Collaboration.
- Creativity.
- Motivation to learn.
Learning Commons is improving its evaluation tools by comparing automated assessments with expert judgments. [5]
Nevertheless, continued validation is needed to determine whether these evaluations remain appropriate across different ages, languages, and cultural contexts.
5.5. Challenge Five: U.S.-Centered Design and International Expansion
Knowledge Graph currently focuses primarily on educational standards from all 50 U.S. states, along with U.S.-based educational research and instructional materials.
This provides important advantages for adoption within the United States.
However, expanding the same infrastructure into countries such as Japan would require more than simply transferring the existing datasets.
For example, deployment in Japan would require consideration of:
- Alignment with Japan's national curriculum standards.
- Subject-specific conceptual structures.
- Japanese-language reading complexity.
- Instructional approaches.
- Differences in grade-level organization.
Although educational AI infrastructure may be technically scalable across countries, educational appropriateness depends on institutional, linguistic, and cultural contexts.
This distinction is especially important when considering international expansion.
5.6. Challenge Six: Long-Term Operations and Financial Sustainability
Learning Commons operates through Learning Commons, Inc., a nonprofit organization, with financial, strategic, and operational support from CZI and activities involving the Learning Commons Donor Advised Fund.
Maintaining open educational infrastructure over time requires continued work, including:
- Updating resources when educational standards change.
- Maintaining and revising instructional datasets.
- Operating and maintaining APIs and evaluation tools.
- Conducting ongoing validation with educational experts.
CZI has also revised its organizational priorities over the past several years. [7]
As a result, important long-term questions concern continued philanthropic support, diversification of funding sources, and operational independence.
This does not mean that Learning Commons has been shown to face a funding shortage.
Rather, these are structural risks that should be considered when evaluating the long-term sustainability of nonprofit infrastructure.
5.7. Summary of the Six Challenges
| Challenge | Current Situation | Key Questions for the Future |
|---|---|---|
| Educational effectiveness | Adoption and implementation examples are expanding. | Independent verification of learning outcomes. |
| Governance | Principles and policies have been published. | Decision-making processes for evaluation standards and opportunities to challenge decisions. |
| Implementation burden | Some schools require additional technical development. | Expansion of implementation support. |
| AI-based evaluation | Evaluation tools are being developed with educational experts. | Validity, fairness, and reproducibility. |
| International expansion | The current emphasis is on U.S. K–12 education. | Adaptation to different educational systems and languages. |
| Sustainability | The nonprofit initiative receives support from CZI. | Long-term financing and operational independence. |
6. Overall Assessment and Future Outlook
Learning Commons occupies a position that differs from both conventional educational AI startups and traditional philanthropic education foundations.
Its most distinctive characteristic is the idea of transforming educational science into shared infrastructure that AI systems can use, while treating the development of that infrastructure as a public-interest activity.
This approach extends the focus of educational AI beyond the performance of individual models and applications.
It places greater emphasis on sharing educational knowledge, evaluating instructional quality, and making effective instructional approaches reusable.
However, understanding its future success requires distinguishing three stages of development.
6.1. Stage One: Making Shared Infrastructure Publicly Available
In September 2026, Learning Commons publicly launched its developer-facing platform. [2]
This milestone has already been achieved.
Resources such as Knowledge Graph, Evaluators, Agent Skills, and Curriculum Sync are now available to support educational AI development.
6.2. Stage Two: Expanding Adoption Across Educational Services
The next stage involves integrating this infrastructure into a wider range of educational products and services.
As of 2026, Learning Commons has been developing collaborations with Anthropic, OpenAI, Canva Education, MagicSchool, TalkingPoints, and other organizations.
School districts have also begun testing applications built with its resources.
Partnership expansion and practical integrations are therefore underway.
6.3. Stage Three: Demonstrating Contributions to Learning Outcomes
Ultimately, the most important question is how much educational AI built using Learning Commons contributes to student learning outcomes.
This stage will require long-term impact assessments and transparent governance.
It will be necessary to investigate whether improvements in educational AI quality translate into better student understanding, knowledge retention, support for teachers' professional expertise, and expanded educational opportunities.
6.4. Three Implications for Educational AI Businesses
The Learning Commons initiative offers three important implications for organizations developing educational AI services.
First, the basis of differentiation in educational AI may change.
If educational datasets and evaluation methods become widely available as shared resources, companies will increasingly need to differentiate themselves through how they combine those resources and the learning experiences they provide to individual students.
Second, evaluating educational effectiveness may itself become an important foundational technology.
The ability to generate questions and teaching materials is not sufficient.
What matters is whether the resulting content aligns with learning objectives and whether its contribution to understanding and student development can be evaluated continuously.
Third, the governance of open educational infrastructure directly affects both competitiveness and trustworthiness.
Who controls datasets and evaluation standards can influence the entire educational AI ecosystem.
Providing resources free of charge does not, by itself, guarantee that the infrastructure serves the public interest.
Integrating educational research and instructional practices into shared AI infrastructure is therefore not only a technical challenge.
It is also a question of how educational quality, accountability, and public value should be maintained.
7. Conclusion
The most important point in understanding Learning Commons is that it should not be viewed simply as another educational AI service launched by CZI.
Its central purpose is to build shared, publicly reusable technological infrastructure for educational AI, drawing on approximately a decade of investment in educational research and development.
This infrastructure brings together educational knowledge, quality evaluation, and curriculum integration.
Following its organizational launch in 2025 and public platform release in 2026, Learning Commons is now expanding its collaborations with companies, research institutions, and schools.
As of November 2026, it is reasonable to describe Learning Commons as an initiative that has made substantial progress in establishing shared infrastructure but still needs to demonstrate educational impact and earn long-term public trust.
The most important question for its future is whether this shared infrastructure can achieve more than making AI development efficient.
Its broader significance will depend on the extent to which it can improve student learning outcomes, support teachers' professional expertise, and expand educational opportunities.
References
[1] Learning Commons, "Scaling Proven Learning Practices," Learning Commons, September 23, 2025. https://learningcommons.org/blog/scaling-proven-learning-practices/
[2] Learning Commons, "AI Platform Launch," Learning Commons, September 22, 2026. https://learningcommons.org/news/ai-platform-launch/
[3] Learning Commons, "Learning Commons," Official Website, 2026. https://learningcommons.org/
[4] Chan Zuckerberg Initiative / Learning Commons, "AI Developer Tools for Education Launch," December 16, 2024. https://learningcommons.org/news/ai-developer-tools-for-education-launch/
[5] Learning Commons, "Learning Commons Platform Launch," Learning Commons, 2026. https://learningcommons.org/blog/learning-commons-platform-launch/
[6] EdSurge, "Learning Commons Launches Open AI Platform for K-12 EdTech," EdSurge, September 22, 2026. https://www.edsurge.com/news/learning-commons-launches-open-ai-platform-for-k-12-edtech
[7] Inside Philanthropy, "CZI Still Funds Education, Just Not How It Used To," Inside Philanthropy, October 20, 2026. https://www.insidephilanthropy.com/home/czi-still-funds-education-just-not-how-it-used-to
[8] Learning Commons, "Data Governance Principles," Learning Commons. https://learningcommons.org/data-governance-principles/