AI Education Fund
Evidence before scale. People in the lead.
A pooled fund for the schools, educators, partners, and families who want to learn what AI can do for DC students, teachers, and leaders. The Fund pays for Evidence Projects in schools, design sprints and ongoing data analysis, cross-sector convenings, learning cohorts, and hands-on AI literacy and vibecoding experiences for the people who run DC schools and the organizations that support them.
Shaped with the Collective’s Steering Committee, Strategic Councils and anchor funders. What you read here is the current design; amounts and dates sharpen as the Fund is capitalized.
AI is already in the classroom. It is up to us what happens next.
Our students are using it. Our teachers are using it. Without people in the lead, AI can make students look like they are learning faster while they are actually learning less: research shows that when students use AI to shortcut problem-solving, grades can improve while real understanding lags. The Fund exists to make sure that when DC students and educators use AI, it sharpens critical thinking and problem solving rather than replacing it. We must be smarter than the tools to use them for long-term benefit.
of DC LEAs have written guidance for staff use of AI.
OSSE, State of AI in DC LEAs, February 2026have written guidance for student use of AI.
OSSE, February 2026high school students report using AI to do their homework, and nearly half of middle schoolers.
RAND, 2026united every DC public and public charter school around evidence-based math instruction through the Capital Math Collective.
Capital Math Collective, 2025To become the first city in the country where all student groups consistently perform at or above the national average in math and reading, with AI as one of many tools that advance that promise rather than undermine it.
The Fund is how the Collective moves from convening to evidence.DC has done this before
- 2021DC launches a statewide high-impact tutoring initiative that now reaches more than 10,000 students a year. Tutored students attend more, with the largest gains for chronically absent students (Stanford), and gain the equivalent of 59 additional instructional days in math (EmpowerK12). OSSE’s initiative Stanford attendance study EmpowerK12 evaluation The 74
- 2025The Capital Math Collective, a $20 million public-private partnership, unites every DC public and public charter school around evidence-based math instruction that accelerates student outcomes. DC Public Education Fund DC PCSB on the launch K-12 Dive
- 2026The Education Recovery Scorecard, a national research project from Harvard, Stanford and Dartmouth, ranks DC first among states in both math and reading growth since 2022. Scorecard: District of Columbia DC case study (PDF) Harvard Graduate School of Education EmpowerK12 on the result
- NextLed by EmpowerK12 — in partnership with DCPS, Charter Alliance, Ed Fund, and Education Forward — the Innovation Fund applies the same playbook to the most consequential shift in education today: shared goals, expert implementation support, community of practitioners, and data that shows what worked.
Five investment areas, one evidence engine
Every dollar goes to people learning what AI can do for students and educators. Choose an area to see what it pays for, who it serves and how we will know it worked.
Evidence Projects in schools
Grants to DC public and public charter schools interested in leveraging AI with staff, teachers and students. A project starts with a need the school/LEA identifies in its own data, tests a tool or a practice against a comparison group, and EK12 publishes what we learned alongside the school.
What it pays for
- School and district leader time to plan, pilot, and reflect on data to continuously improve
- Tool licenses and implementation coaching for the pilot term
- Experiment and quasi-experimental evaluations with the DC Education Research Collaborative
- Educator and student participation in co-design and evaluation
How we know it worked
Student learning and sense of belonging indicators, focusing on priority groups like economically disadvantaged students, students with disabilities, and multilingual learners. Ongoing data analysis will be used to identify what's happening, working, and moving the needle. Results will be made pubic regardless of whether they are promising or not.
Design sprints and ongoing data analysis
Support for AI implementation in ways EK12’s that the successful math acceleration grants work: a short, facilitated design sprint that turns a need into a plan with a hypothesis and a measure, then data analysis through the year so teams adjust while it still matters, not after the school year ends.
What it pays for
- A design sprint for every funded project, facilitated by CitySchools Collaborative
- Monthly implementation data reviews with EK12 analysts
- Comparison groups drawn from DC’s shared data
- Shared measures and guidance in the EK12 Bold Improvement Portal
How we know it worked
Whether teams reach ambitious milestones they set in collaboration with EK12, and implememntation improvements made based on data and tracked for longitudinal impact analyses.
Cross-sector convenings
Gatherings that bring DC Public Schools, public charter schools, government, funders, nonprofits, industry partners and families to one table to share bright spots from DC classrooms and updates on the latest research on AI in schools. No school should have to solve the same problem alone.
What it pays for
- Bright-spot sessions where funded teams show their work
- Research briefings with national and local researchers
- The annual AI Education Leadership Summit and Funders Forum
- Exchange with aligned funds in New Orleans and Tennessee
How we know it worked
Who came, what they took back to their schools, and which practices spread from one school to the next.
Grantees present their results in the open, wins and misses alike.
Cohorts for executives, school leaders, teachers and nonprofit professionals
Paid places in structured cohorts for each stakeholder group: LEA executives, school leaders, teachers, and professionals at the nonprofit organizations that support DC schools. A cohort is a semester of learning with peers, anchored in change management and evidence, not a one-time workshop.
What it pays for
- Cohort facilitation and materials
- Participant stipends and substitute coverage where needed
- Coaching between sessions
- Places in national communities of practice
How we know it worked
Leadership capacity, teacher engagement and student agency with AI, benchmarked against other cities so DC can see where it stands.
Training, training, training, and not one and done.
AI literacy training and vibecoding experiences
Design and delivery of hands-on AI literacy training and vibecoding experiences for the same four groups: how AI works, where it fails, how to direct it, and how to build a working tool for your own school or organization by describing what you need. The aim is people who understand AI well enough to harness it, so the education sector becomes more effective and more efficient in ways that complement what we already know works for students.
What it pays for
- Curriculum designed with DC educators
- Half-day and multi-week experiences, in person and online
- Build sessions where teams leave with a working prototype
- Student data privacy and safety built into every session
How we know it worked
Skills demonstrated, tools built and still in use a semester later, and hours reclaimed for the human parts of the job.
It has always been true that you have to be smarter than the tool for it to help. AI is no different.
Built for the people who run DC schools, and the organizations that support them
Choose your role to see what the Fund offers and what it asks in return.
For LEA executives
Superintendents, CEOs, chief academic officers and network leaders
What the Fund offers you
- An executive cohort on leading with AI: vision, policy, procurement and change management
- Evidence Project grants for schools in your network, sized to the need
- Comparison-group analysis across DC that no single LEA can run alone
- A seat at the convenings and in a standing network with New Orleans and Tennessee
What we ask of you
- Name the need in your own data before the tool
- Open your doors and share results with other DC schools
- Protect student data and use only tools that meet legal standards
For school leaders
Principals, assistant principals and instructional leaders
What the Fund offers you
- A grant and a design sprint that turn a real problem into a plan
- A school-leader cohort with peers facing the same decisions
- Quarterly data reviews so you adjust mid-year, not after
- Bright-spot showcases that put your team’s work in front of the city
What we ask of you
- Protect teacher time for the pilot
- Bring families and students into design and evaluation
- Report what worked and what did not
For teachers
Classroom teachers, coaches and teacher leaders
What the Fund offers you
- A teacher cohort with a stipend, and coaching between sessions
- AI literacy and vibecoding sessions built with DC teachers
- Time-back use cases tested with evidence: planning, feedback, family communication
- A voice in what the Fund supports next, through the Strategic Councils
What we ask of you
- Pilot honestly and keep the relationship with students at the center
- Tell us where the tool helped and where it got in the way
For education support organizations
Nonprofits and intermediaries that serve DC schools and students
What the Fund offers you
- Cohort places for your staff
- Vibecoding experiences to build tools for your own operations
- Partnership on Evidence Projects as the practice partner to a school
- Convenings where your bright spots reach school leaders
What we ask of you
- Practitioner-led projects: educators lead, technology supports
- Share what you build so other organizations can use it
For students and families
The people every project is for
What the Fund offers you
- A seat in the design and evaluation of every funded project
- Family AI literacy nights and plain-language guides
- Student-facing experiences, from AI literacy to building for peers and schools
- Answers to one question, in public: what changed because students and families were in the room?
What we ask of you
- Tell us what you see at home and in class
- Hold the Fund to its people-first promise
From a need in the data to a result the whole city can use
Five steps, each one a habit borrowed from work that already moved outcomes in DC.
Name the need in your data
A school or partner names an opportunity in its own student data: a grade level, a subject, a group of students. No tool yet. Several schools elsewhere bought licenses for everyone and could not say what need they met. DC starts the other way around.
Form the team
A school, EK12 plus CitySchools as sprint and data partners, a researcher through the DC Education Research Collaborative and, where it helps, a developer or a nonprofit partner. Educators and education leaders lead; technology partners support but do not drive.
Run a design sprint
Four weeks to a plan, facilitated by CitySchools: the question, the comparison group, the measures students and teachers already produce, the guardrails for student data, and what the team will do if the answer is no. The same method EmpowerK12 used with 32 charter networks on high-impact tutoring.
Pilot with a comparison group
A term or a school year of implementation with quarterly implementation data reviews run by EmpowerK12, so the team adjusts while it still matters. Short pilots are judged as learning, not held to multi-year outcome standards.
Share what you learned
A public write-up with the DC Education Research Collaborative, a bright-spot session, and results shared with aligned funds in other cities. A pilot that shows an idea does not work saves every other school the cost of finding out.
What every grant commits to
Need before tool
Applicants name the opportunity in their own student data before any tool enters the conversation.
Families and students in evaluation
Every project includes educators, families and students in design and evaluation before broader implementation.
Open doors
Every grantee hosts visits, shares data and lessons, and contributes exemplars to the DC AI Practice and Resource Bank.
Short pilots judged as learning
A six-month pilot that proves an idea does not work is still a result. Teams are not locked in beyond the pilot.
EmpowerK12 runs the Fund, with the Collective at the table and researchers alongside
EmpowerK12, the Collective’s co-lead and DC’s education data partner, runs the Fund. It supports every design sprint and the ongoing implementation data analysis, the way it supports DC’s math acceleration grants today, and it works with the DC Education Research Collaborative to do the research, so every result is independent, comparable and public.
EmpowerK12
Holds the Fund on behalf of the Collective and staffs it: the open calls, the grant agreements, the fund committee, reporting to funders, and the data spine every project shares. The Fund lives where DC’s data-sharing agreements, the Bold Improvement Portal and the research partnerships already live.
EmpowerK12 with CitySchools Collaborative
Facilitates every design sprint and runs the ongoing implementation data analysis: quarterly reviews with each team, shared measures in the Bold Improvement Portal, and comparison groups drawn from DC’s data under existing agreements.
DC Education Research Collaborative with EmpowerK12
The research-practice partnership housed at the Urban Institute that provides independent research on DC’s public and public charter schools. Together researchers and educators design each study, evaluate the Evidence Projects, run the cross-site studies with New Orleans and Tennessee, and publish findings.
The Collective and its funders
The Steering Committee and the Strategic Councils shape what the Fund invests in. A fund committee with Collective, family and student seats sets priorities each year, and anchor funders meet regularly as the strategy and pressure-testing group.
A four-year acceleration fund, then a planned wind-down
puThe Fund is built to move fast and use data in rapid improvement cycles, scaling what works and discontinuing what doesn't show promise. Most of the fund will be spent by the 2029-30 school year with a wind-down year to ensure LEAs have time to leverage per pupil funds to continue the right work and publish final research results.
The Fund is designed in collaboration with the Steering Committee, Strategic Councils, and the October Funders Forum. Anchor funders commit and a fund committee is at the table by winter.
The bars show the planned shape of the Fund’s activity by school year, not amounts. Dates are targets and firm up as funding is confirmed.
Year 1, 2026-27: launch
The first open call goes out. Selected teams start with design sprints, and the first Evidence Projects and cohorts begin. An adoption rubric and a baseline picture of AI in DC schools are published.
Year 2, 2027-28: build
A second call grows the portfolio. The first evaluations publish and the first tool or practice is recommended for citywide adoption. Cohorts double, and the aligned funds meet in DC at the 2027 Summit.
Year 3, 2028-29: peak
The Fund’s largest year: the most new projects, the first cross-site study with New Orleans and Tennessee, and DC’s first replicated result.
Year 4, 2029-30: scale
The final round of new projects. Most of the Fund is spent by the end of this year. Every completed evaluation publishes, tools and practices with evidence are recommended citywide, and the 2030 State of AI in DC Schools report goes out.
Wind-down, 2030-31
No new grants. The last evaluations publish, DC schools and districts carryshared measures forward, and schools carry proven tools and practices on their own per-pupil budgets.
The Collective’s principles, applied to money
The Fund uses the same eight principles the Collective set for all of its work. Each one is a test a project has to pass.
Ground-up stakeholder engagement
Educators, families and students shape priorities through continuous, structured collaboration.
Evidence-driven
Decisions rest on rigorous evidence, not hype or claims. DC Public Schools and public charter schools pilot together, measure impact, and share what works and what does not.
People-first
AI supports and empowers educators and students. It does not do the work for them, and it fosters creativity and critical thinking rather than automating them away.
Equitable access and inclusivity
Tools reach every student, narrow the digital divide, avoid bias and meet the needs of diverse learners across the District.
Ethical usage and data privacy
Strong data security practices, and only AI technologies that meet legal standards for student data privacy.
AI literacy and agency
Teachers and students understand how AI tools work, recognize their limits, and learn to use them safely and effectively.
Transparency and accountability
Schools say when and where AI is part of learning, people verify AI outputs for accuracy, and developers are held to ethical standards.
Pedagogical purpose
AI serves learning goals, not the other way around. The relationship between students and teachers comes first; the support tools come after.
Borrowed from a city that moved first
In 2025, New Orleans launched a five-year, locally grown fund to help its public schools make strategic, instructionally sound decisions about AI. Its first year reached leaders of schools enrolling 60 percent of the city’s students and moved into classrooms with tools tied to real academic needs. DC’s Fund borrows its strongest features and adds the one thing a citywide evidence engine needs most: a shared data spine.
What we borrow
- Need before tool: leaders name the academic opportunity before any tool is chosen
- A stated sunset, so no strategy conversation is about self-preservation: for DC, four years of acceleration and a wind-down year in 2030-31
- Funders who coordinate their grantmaking and meet often, rather than a distant board
- Leader capacity first, then implementation, then redesign
- Every grant shares openly and opens its doors to other schools
- Measures that matter: is learning improving, are teachers staying, is capacity lasting
What DC adds
- EmpowerK12 runs the Fund, supports every sprint and the data analysis, and does the research with the DC Education Research Collaborative
- Comparison groups drawn from EmpowerK12’s data-sharing agreements with more than 40 DC LEAs
- Shared measures in the Bold Improvement Portal, EmpowerK12’s data platform for DC schools
- Design sprints and quarterly data analysis, the method behind DC’s math acceleration grants
- Every project publishes its result, yes or no
- Cohorts and AI literacy for executives, school leaders, teachers and nonprofit professionals, not school leaders alone
- A standing network with the New Orleans and Tennessee funds: common measures, coordinated calls and shared studies
A five-year fund with a stated sunset.
of New Orleans public school students attend schools taking part in 2026-27.
student writing assignments received standards-aligned AI feedback at KIPP New Orleans in 2025-26.
AI Innovation Index measures grew across New Orleans partner schools in year one.
Tennessee SCORE’s AI Innovation Fund made nine grants of up to $50,000 in 2026, each practitioner-led, measurable and safe for students. DC’s rapid pilots follow that shape; its multi-year implementation grants follow New Orleans.
Three questions guide every dollar
Borrowed from New Orleans, answered with DC data by EmpowerK12 and the DC Education Research Collaborative.
Is student learning improving?
Grade-level performance and growth on the measures DC already uses, from classroom assessments to CAPE and NAEP, reported for priority groups first.
Are teachers getting time back, and choosing to stay?
Hours reclaimed for the human parts of the job, teacher engagement and satisfaction, and retention at participating schools.
Are schools building capacity that lasts?
Leadership capacity and AI literacy benchmarked against other cities, tools still in use a year later, and practices that spread from one school to the next.
Results publish twice a year, on this site and with the aligned funds, whether the news is good or not.
Frequently asked
Who can apply?
DC public and public charter schools and LEAs, and nonprofit organizations serving DC students and young people through age 24, usually in partnership with a school. Public charter LEAs are likely to lead the first round, with DC Public Schools joining as its policy environment allows.
Do we need to have a tool in mind?
No. Need before tool: applicants name the opportunity in their own student data first. The design sprint then matches the need to a practice, a tool, or neither.
How big are the grants, and how long do they run?
Two shapes are planned. Rapid pilots run six to twelve months and answer one question. Implementation grants run up to three years and are front-loaded, so a school absorbs the cost as its own capacity grows. Amounts publish with the first open call.
How long does the Fund itself last?
Four school years of acceleration, 2026-27 through 2029-30, with most of the Fund spent by the end of 2029-30. The 2030-31 school year is a wind-down year: no new grants, the last evaluations published, and schools carrying what worked on their own budgets. Nothing is funded past the wind-down year, and every multi-year grant is sized to end inside it.
What does a grantee commit to?
Families and students in design and evaluation. Open doors for other DC schools. Shared measures and a public write-up. Student data privacy, with only tools that meet legal standards.
Is a pilot that fails a failure?
No. Short pilots are judged as learning. A six-month pilot that shows an idea does not work saves every other school in the city the cost of finding out, and teams are not locked in beyond the pilot.
Where does the Fund live, and who decides?
EmpowerK12, the Collective’s co-lead and data partner, runs the Fund on behalf of the Collective: it holds the money, runs the open calls and the grant agreements, and keeps the data-sharing agreements, the Bold Improvement Portal and the research partnerships in one place. A fund committee with Collective, family and student seats sets priorities, and anchor funders meet regularly as the strategy and pressure-testing group.
Who does the research?
EmpowerK12 works with the DC Education Research Collaborative, the research-practice partnership housed at the Urban Institute that provides independent research on DC’s public schools. EmpowerK12 facilitates the design sprints and runs the ongoing implementation data analysis; the Collaborative and EmpowerK12 together design each study, evaluate the Evidence Projects and publish what they find.
How is student data protected?
Comparison groups come from data EmpowerK12 already holds under agreements with DC LEAs and are analyzed under those agreements; small groups are suppressed in anything published. Only tools that meet legal standards for student data privacy enter a funded project.
How can funders take part?
Multi-year gifts pooled at EmpowerK12, a committee seat for anchor funders, co-funding of cross-site studies with the aligned funds, and capacity as well as cash: space, engineering time, licenses and research partners.
How does this connect to the Capital Math Collective?
Same playbook. The Math Collective united DC Public Schools, public charter schools, OSSE and philanthropy around common goals, used design sprints and ongoing data analysis, and helped make DC the most improved state in math since the pandemic. The Innovation Fund applies that discipline to AI.
Help build DC’s evidence engine
Fund it, pilot it or shape it. Every one of these starts with a conversation.
For funders
Seed the Fund with a multi-year gift, take a committee seat, or bring capacity: space, engineering time, licenses and research partners.
Talk with usFor schools and LEAs
Tell us the need you see in your data. We will let you know when the first open call opens and how the design sprint works.
Express interestFor partners and councils
Join a Strategic Council or the Steering Committee and shape what the Fund supports next.
How we are leadingStart with an email
Write to the Collective at aicollective@empowerk12.org and someone from the team will follow up.