The events at Roosevelt High School in Minneapolis on January 7 were a nightmare. Armed Border Patrol agents arrived during dismissal, tackled people on school grounds, handcuffed two staff members and appeared to deploy chemical agents — including on students. The district canceled classes for the rest of the week. Earlier that day, a woman was killed by an ICE agent just blocks away.
While educators scramble to respond to this crisis and others like it across the country, education philanthropy is making billion-dollar bets on a different future: Major foundations are pouring unprecedented resources into artificial intelligence in education.
These are powerful technologies. But at this exact moment, 26% of school-age children come from immigrant-origin families. Those students are watching what happened at Roosevelt and wondering if their school could be next, and so are the classmates sitting beside them. Meanwhile, the philanthropic sector is treating technical capacity as education’s primary gap.
This is a fundamental confusion of tools with outcomes. AI answers “how” but not “what for.”
The philanthropic disconnect
The scale of investment in educational AI is staggering. The Gates Foundation, Chan Zuckerberg Initiative and others are directing billions toward these technologies. Meanwhile, funding for the work that actually builds belonging — professional development for educators, curriculum transformation, student-led after school initiatives — is quietly disappearing. As immigration and belonging have become politically contested, many funders are retreating to what feels safer: technical solutions that promise to improve outcomes without wading into cultural battles.
But this retreat is happening at precisely the wrong moment. Recent UCLA research surveying over 600 high school principals nationally reveals that 70% report students from immigrant families expressing fear about their safety. Nearly two-thirds report students missing school due to immigration policies and rhetoric. More than a third report bullying and harassment directed at immigrant-origin students. And 58% report immigrant families leaving their communities entirely.
These patterns appear everywhere: in Tennessee, Wisconsin, Nebraska and Minnesota, not just in border states. A Wisconsin principal overheard a student tell a classmate, “Just make sure you have your ID with you.” A Texas principal described students suddenly telling immigrant-origin classmates whom they had known since third grade to “go home.”
This crisis was predictable. Robert Putnam’s 2007 “E Pluribus Unum” article documented how diversity can initially decrease social trust. Ryan Enos showed how even brief exposure to difference shifted attitudes. Jennifer Richeson and Maureen Craig demonstrated how demographic projections alone trigger anxiety among white Americans experiencing “majority status threat.”
But the research also showed contact works under the right conditions: sustained interaction across difference, equal status, common goals and institutional support. The problem isn’t diversity — it’s that we’ve systematically eroded the settings where positive contact happens. COVID accelerated this collapse. The isolation we experienced, combined with streaming video replacing shared public entertainment, meant that Americans increasingly encountered difference through screens rather than in person. The brief surge of support for diversity after George Floyd’s murder couldn’t hold without sustained infrastructure to support it.
Schools as civic infrastructure
Schools are the last universal gathering place in American life. Unlike parks or museums, schools already have the conditions that can bridge difference and build belonging: They bring together young people who might otherwise never meet, sustain those relationships over years, structure cooperation toward shared goals, and provide institutional support for navigating conflict and connection. Decades of research on intergroup contact has illuminated practices that leverage these conditions to reduce prejudice and foster mutual understanding. But these practices have been no one’s priority. They need to be taught to educators, implemented in classrooms and supported by school leadership.
If that happened — if belonging practices were systematically embedded in school culture — the potential leverage is extraordinary: 56 million young people, in the one institution that still brings Americans together across every line of difference. Ambitious, yes — but the infrastructure exists. The question is whether we invest in it.
A recent UCLA survey from John Rogers and Joe Kahne shows principals doing what they can to preserve these spaces and respond to escalating threats. Seventy-eight percent have created emergency response plans for potential ICE visits. Nearly half have protocols for supporting students if parents are detained. Educators are connecting families with legal services, creating safe spaces for difficult conversations.
But emergency protocols cannot prevent students from absorbing anti-immigrant rhetoric and directing it at classmates. Legal services cannot rebuild belonging after someone a young student trusted tells them they don’t belong. And no crisis planning can undo what students at Roosevelt witnessed.
What’s needed is long-term investment in transformation: professional development so that every teacher sees belonging as their responsibility, curriculum that weaves migration through a range of subjects, student-led initiatives where young people become architects of inclusive school cultures. This prevents crisis, rather than just responding to it.
This work complements but differs from traditional civic education. Civic ed funders have an important role here — but preparing students for pluralistic democracy requires more than civic knowledge measured by standard indicators.
Teaching constitutional literacy and government structures is essential, but it doesn’t prepare communities to actually navigate difference and build belonging across it. And when civic education assumes citizenship as the only pathway to participation, we’re teaching a model that federal policy is increasingly rendering impossible and irrelevant for many of the immigrant-origin youth who comprise a quarter of our school-age population. The opportunity for civic education funders is to recognize that civic capacity for pluralism is itself a form of civic knowledge — one that schools desperately need resources to develop.
And this isn’t just about immigrant-origin students. Design schools that genuinely work for students navigating multiple identities and languages — where the most vulnerable truly belong — and everyone benefits. It’s the “curb cut effect”: design for the margins and you improve the whole system.
A different kind of bet
Concretely, this means supporting the professional development of educators in ways that are responsive to local demographic realities. A school in rural Wisconsin navigating demographic change requires different approaches than an urban school in Texas with long-established immigrant communities. What both need are educators who understand belonging frameworks and can translate them to their particular contexts — educators whose instinct is to learn about their students’ lives and stories, to build authentic relationships across difference, and to advocate when those students face barriers inside or outside school walls. They also need administrators who support teaching about migration as central to American identity while addressing the specific questions and anxieties emerging in their hallways.
This is what organizations like Re-Imagining Migration are doing through regional fellowship networks, supporting sustained coaching and communities of practice where educators learn with and from each other, not one-time workshops or off-the-shelf curriculum packages. This capacity-building work exists and is reaching educators at scale, but it requires funders willing to make multi-year commitments during a politically contested moment rather than waiting for safer conditions.
The cost of waiting
Education philanthropy has a pattern: retreat when issues become politically contested, return when “conditions improve.” We saw this with civics education, defunded for decades. We’re seeing it with diversity and belonging work today.
The stakes are different now. Demographics aren’t reversible — immigrant-origin students make up 26% of the student population and that percentage is growing. Meanwhile, civic infrastructure has already been weakened. Schools are under simultaneous political pressures.
If philanthropy waits for political winds to shift, we’ll have raised a generation of children who learned that belonging is provisional, that their schools weren’t designed for them. That’s not a technical problem AI can solve later. That’s a civic failure that compounds across generations.
The question isn’t whether AI and other ed tech solutions have value. The question is whether, at this moment of democratic reckoning, we’re investing in the civic outcomes that actually matter. Philanthropy must sustain — not retreat from — investment in schools as civic infrastructure for belonging. This means funding educator networks, curriculum resources that center migration as an American story, student voice initiatives, and sustained partnerships that help schools become genuine communities rather than content-delivery systems.
The market won’t solve this. Policy won’t solve this until we’ve built civic capacity for it. If philanthropy abandons this work now, we risk losing a generation’s faith in whether American institutions are designed for them.
That’s the choice philanthropists are making with their 2026 budget allocations. Fund technical infrastructure, or fund civic infrastructure. One teaches skills. The other builds democracy.
Adam Strom is Co-Founder and Executive Director of Re-Imagining Migration, a national organization that transforms K-12 education into civic infrastructure for belonging.
Note (2/4/26): A previous version of this piece mentioned Schmidt Sciences as one funder of educational AI. Schmidt Sciences does not currently back the kind of educational AI work referred to here.
Related Inside Philanthropy Resources:
For Subscribers Only
