When districts and state agencies make important decisions, if they consult their communities at all, they usually hire outside experts to do it. Armed with surveys and interview protocols, these consultants go into the community, collect feedback, synthesize it, and report back. The community is a data source; the district a receiver of findings. The consultants are the real meaning-makers.

Last week, we shared a different approach and raised empathy interviews as a key tool education leaders can use to go beyond superficial stakeholder engagement, gaining a more authentic picture of the real-life experiences of the people they serve. In our Guide to Empathy Interviewing, we asserted that leaders who conduct empathy interviews will be able to “create systems that address real needs, build broad support, and foster shared accountability for important work ahead.”

But that’s only half-true. Results don’t come from just using the tool. We’ve learned that how you use it, specifically, who conducts the interviews, and what you do with the data afterward, that matters most.

Here’s the other half of our approach.

Who Conducts Empathy Interviews?

Earlier this week, CIE supported district leaders in Burlington School District, Vermont, to design empathy interview questions to ascertain what different groups of people think of the district’s current work and priorities. The effort will help inform the priorities, action areas, and metrics of the district’s next strategic plan.

On its face, there’s not that much different here than any fact-finding process. But here’s the catch: we (the consultants) support the overall process, but Burlington leaders and members of their community will be the ones conducting the interviews. A district leader intently listening to a teacher or student. A parent interviewing another parent. A student interviewing a peer.

A lot changes when you shift the driver’s seat from outsiders to members of the community. When a consultant interviews a community member, that person is being studied. When a neighbor interviews a neighbor, that person is being heard by someone who already shares some part of their context, and who is going to carry their story back to a group that will make decisions informed by what they hear.

This changes whose knowledge gets centered and what questions even get asked. A student who has experienced a particular kind of exclusion knows what follow-up to ask that no outside researcher would think of. A parent who has navigated a system not built for their child understands immediately when another parent is describing a similar experience and knows how to inquire with discernment and respect.

We don’t prioritize interview uniformity. The interview adapts to the interviewee as it unfolds, shaped by the context that they share.

One other thing that happens when district leaders and members of the community conduct the interviews: the interviewers themselves are changed. Time and time again, the people we have trained to conduct empathy interviews have called their experiences “eye opening” and “incredibly powerful.” The act of sitting with someone else and deeply listening to how they experience the world impacts how we perceive the world, too. And that changes everything about what the interviews can produce.

What You Do With It Matters

We have found that the interviews are critical, but they really are only the beginning.

Once the coalition of district leaders and community interviewers brings their anonymized interviews back to the group, the work of sense-making begins. And this is where something important happens that gets lost if consultants do all the work.

Every coalition member reads synopses of every interview. Not just the ones they conducted, but from a broader range of perspectives beyond their own. The breadth of what the coalition absorbs together becomes the foundation for a shared reality. The synthesis they create together is not a compromise between competing interests, but a genuinely expanded picture of what’s true for the community the district serves. And from that shared reality, the coalition can begin to identify root causes together: not from the individual wounds and victories that each person entered with, but what the aggregation of all those stories points toward. Forging a shared reality is challenging and exciting facilitation work, but worth the effort.

When members of the community aren’t just interviewed, but are invited to make sense of the data and shape what comes next, they become participants. Participants become co-creators. Co-creators become stewards of a plan they helped build. They advocate for it, and they also hold the district accountable for upholding it.

The result is not just a better strategic plan. It is the start of a more reciprocal relationship between a community and its schools. One where community members feel they have a stake in what happens with their district, because they were “in the room where it happens.” They were there when the questions were asked, and there again when the answers started to become decisions.

When we ask who gets to conduct empathy interviews, we’re really asking something deeper: Whose knowledge counts? Who gets to be the author of our shared understanding? Who are we trusting to carry the community’s stories into the room where decisions get made?

Those are not neutral choices. They communicate something. And in Burlington, the answer is: the students, caregivers, educators, and community members that have a stake in the future of its young people. Trained, supported, and empowered coalition members who will go into their own networks, have honest conversations, and bring back what they find. These finding are not just data points, they stories that will reshape how the district understands itself.

That’s what empathy interviews make possible. That’s why it matters who does them, and what happens next.

Ready to get started? Download our Guide to Empathy Interviewing here. If you’ve used empathy interviewing in education planning processes before, or want to try, we’d love to hear your thoughts!

What do you picture when you hear “stakeholder engagement?” A warm scene of solidarity and shared purpose, or a cold image of lip service and performative posturing? Too often, we’ve seen education leaders invite feedback from stakeholders in superficial ways that lead to the latter impression.

Many of us are working to flip that script, using terms like co-creation to describe a deeper kind of stakeholder engagement–one that doesn’t just ask students, families, and community members for their opinions (with no promise of action), but includes them in the very rooms where decisions are made. We’ve seen investments in co-creation result in systems and strategies that better address root problems and create broad community buy-in.

Understandably, not every leader is in a position to commit to a full, end-to-end co-creation process. Doing so requires time, structure, and thoughtful facilitation. But in our experience, co-creation always starts with one step, and it happens to be one that any education leader can commit to with immediate benefits.

That small but mighty step is the empathy interview.

An empathy interview is a one-on-one, open-ended conversation designed to surface not opinions or preferences, but beliefs, emotions, and needs. Not a sterile focus group or a town hall where the loudest opinions reign, an empathy interview is a conversation between two people, in which one of them — the interviewer — is trying as hard as possible to understand the lived experience of the other. The questions are open. The follow-ups are curious, not leading. The goal is to get to the story underneath the answer.

In 2021, we supported Burlington School District to co-create its next strategic plan with a coalition of community representatives. We trained district leaders, teachers, parents, students, and community members on the coalition, who conducted more than 75 interviews by going into their own networks, talking with neighbors, family members, and people who would never show up to a school board meeting or a town hall.

The results were striking. The process allowed district leadership and the coalition to connect with new people in order to bring in insights that are typically unavailable, and to see patterns across those experiences. For example, one interviewer said,

“It was really eye opening to [hear others] sharing painful stories about how students were not being served and their frustrations. This was profound for me… I had a fuller picture from it all.”

A district leader reflected on what the process made possible for them personally:

“The empathy interviews were incredibly powerful. [My children] are white, cis-gendered, able-bodied students, so I had a very limited perspective. But the empathy interviews, and hearing all the different stories… gave me pause, and really allowed me to expand my vision and my view.”

This is what empathy interviews do that focus groups and surveys can’t. They reach the people and the stories that don’t make it into the room on their own. And they do it through relationships and the simple, radical act of one person saying to another: Tell me what your experience has been. I’m here to listen.

By bringing all of their interview experiences together, the Burlington coalition identified patterns and common issues that became the focus of the district’s strategies and actions. We’re now gearing up to support Burlington’s next round of strategic planning, leveraging the district’s capacity to conduct empathy interviews and build from their community’s collective experiences.

Meanwhile, we created this Guide to Empathy Interviews so others can learn more about the process and find out how to get started.

In our next post, we’ll describe how we use empathy interviews, including who should conduct the interviews (hint: not an external consultant) and what to do with the information. These decisions can mean the difference between a process that gets “interesting feedback” and one that transforms systems and builds community trust. Stay tuned!

Download the CIE Empathy Interview Guide here!

We say “it takes a village to raise a child,” but when was the last time anyone asked what it actually takes to build one?

That question has been at the center of a partnership between the Fox Chapel Area School District and the Center for Innovation in Education for some time now. We are proud to announce that we have been selected as a recipient of a 2026 National Moonshot Grant from Remake Learning to explore this question. We are one of 15 projects nationwide chosen to test bold, future-facing ideas in teaching and learning.

Our project is called Building the Village: Schools as Infrastructure for Community Prosperity. The grant is supported by The Grable Foundation, Siegel Family Endowment, Richard King Mellon Foundation, The Heinz Endowments, and The Buhl Foundation. We’re honored to be in the company of the other Moonshot grantees, and genuinely excited about what we’re being given the chance to try.

What we noticed and why it bothered us

Look honestly at how community efforts typically engage with one another and you’ll find what we started calling “parallel play” - something akin to watching two young children playing next to one another, each unaware of what the other is doing. We often see good people, good programs, and good intentions running side by side, but rarely adding up to something larger than themselves. A walking group here, a service project there, a community garden that are all beautiful and good but often falling short of cumulative impact.

At the same time, we’ve watched schools positioned as institutions for young people rather than as critical infrastructure of community life. We often act as if the school district’s work ends at the building’s edge, and that young people are leaders-in-training rather than leaders right now.

We don’t believe that. And the research on social capital and shared cooperative governance suggests that communities thrive when institutions are genuinely integrated, when the work is coordinated and cumulative, and when people of all ages contribute to shared goals. Instead of parallel play we are envisioning cooperative village-building.

But then we wondered, what’s a shared goal that might be energizing for an entire community and that would inspire unprecedented coordination between organizations? This led to the other animating question behind this project: What would it take for Fox Chapel Area to become a Blue Zone?

What we’re building

Blue Zones are communities where people measurably live longer, healthier, more connected lives. Not because of individual willpower, but because of the environments and social structures around them. That’s the target we’re orienting toward.

Over the next 12 months, the Fox Chapel Area School District and CIE will co-create a Community Impact Council, a coordinating body that brings together the school district, municipal partners, community organizations, and students. The Council will work towards building two core products: a Community Prosperity Roadmap and a shared accountability scorecard.

The Roadmap isn’t just a vision document. It’s a tool for coordinating distributed innovation across intergenerational teams that might be working on things as disparate as walking groups, cooking circles, health literacy initiatives, and community gardens. The Roadmap will help ensure that what might otherwise be isolated efforts compound into measurable community progress. The Community Impact Council is the structure that makes that coordination possible. It is the body that holds the teams accountable to each other and to the shared direction, and that ensures that the work adds up rather than runs in parallel.

The scorecard is how we know if it’s working. Rather than importing someone else’s metrics, the Roadmap process will surface the leading indicators that matter most here by gathering the early, trackable signals that the Fox Chapel Area School District community is actually moving toward Blue Zone conditions. Those indicators become a shared accountability tool that both the school board and the local governments use to ask the same question at the same table: How are all of us, across generations, actually doing?

The work unfolds in four phases:

In the Discovery and Roadmap Co-Creation phase, we’ll begin with empathy interviews. These participant-led will center on conversations with families, community members, and organizational leaders about what prosperity actually means here. We’ll map assets, identify bright spots, and co-design the Council’s structure. The Roadmap emerges from that process, not before it.

In the Launch of Intergenerational Teams phase, we’ll form three to four teams around the Roadmap’s priorities, which might include things like walking groups, cooking circles, health literacy initiatives, community gardens. We will give these teams the coordination infrastructure to turn isolated contributions into cumulative progress.

In the Scaling and Governance Integration phase, we’ll formalize the joint governance routines such as school board and local council sessions where Roadmap data is reviewed together, where accountability is shared rather than siloed.

And in the Reflection and Replication phase, we’ll document what we learned honestly, including where coordination broke down, so that other districts and municipalities can adapt the model for their own contexts.

What we hope this becomes

From Mary Catherine’s vantage point as superintendent, this project represents something she’s been working toward for years: a version of school-community partnership where the district isn’t just a venue for programs, but a genuine engine of shared prosperity. Fox Chapel Area students already demonstrate extraordinary leadership capacity. This initiative asks: what happens when we give that capacity a real coordination role on a wider scale, not as a future aspiration, but now?

From Doannie’s vantage point at CIE, this is about proving something structural, rather than a program that depends on heroic individuals to sustain it. We want to go beyond community engagement as performance towards infrastructure. In other words, replicable, transferable blueprints for building villages whether you’re in a suburban district, a rural township, or an urban neighborhood with different assets but the same coordination gap.

What we both believe is this: the structures of community already exist in the schools, local governments, civic organizations, faith communities, businesses within them. What’s missing isn’t goodwill. It’s coherence. A shared roadmap. A coordinating body with real accountability. Students treated as essential contributors, not as future citizens waiting in the wings.

That’s the village we’re trying to build in the Fox Chapel Area School District community. And if the model works here, the playbook belongs to everyone.

Mary Catherine Reljac is the superintendent of the Fox Chapel Area School District.

This project is supported by a National Moonshot Grant from Remake Learning. National Moonshot Grants are made possible by The Grable Foundation, Siegel Family Endowment, Richard King Mellon Foundation, The Heinz Endowments, and The Buhl Foundation.

Pull the screens out, but make sure students can use AI.

The contradiction in these directives shows we don’t know enough about the technology we’re trying to regulate. But even more concerning is that both campaigns are driven by adults, acting on behalf of youth, almost entirely without them.

That should give us pause. Previously, we wrote about the costly disconnect between adult priorities and students’ experiences with generative AI. Youth are missing from the anti-screens evidence, too: a recent analysis found that of the 26 experimental studies cited by Jonathan Haidt in The Anxious Generation — the book fueling cell phone bans — only one study involved actual K-12 youth.

Early data shows what happens when policy isn’t grounded in youth reality: in the largest study of restrictive phone-storage pouches, NBER found a 16% spike in student suspensions in year one but virtually no improvement in the test scores, attendance, classroom attention, and online bullying that the policy was supposed to improve. Other studies of cell phone bans show similar patterns of shortcomings and resistance.

We shouldn’t be surprised. Too often, top-down, restrictive policies miss important nuance, fail to generate buy-in, and seed unintended consequences. In contrast, policies co-created with the people they impact are far more likely to address real issues and build lasting ownership.

Listening to Student Experiences

To understand how youth actually experience technology, the Center for Innovation in Education studied cell phone experiences in a large urban-to-suburban school district. Required by state law to develop a cell phone policy, the district’s initial community surveys revealed familiar splits: more teachers supported restrictive policies than did students, while parents were mixed. But to find a way forward, they sought to better understand why people feel the way they do.

For that, we used SenseMaker, a method that gathers personal stories and surfaces patterns across them. Instead of opinions, we asked for people to share a real moment that could only happen because of a phone, smartwatch, or earbud, and to tell us what it means. The approach maps underlying values more authentically than a survey, and widens the lens beyond the classroom. We collected 932 first-person accounts of life with technology, 205 from students.

What Their Stories Revealed

Like the district’s initial survey, national narratives position student, parent, and teacher values at odds: students want social stimulus, parents want safety, and teachers want focus. Our story data suggested a different picture.

Students, educators, and parents expressed the same underlying values in roughly equal measure: feeling safe and able to get help, connected to people who matter, and in control of getting things done. The divide wasn’t in what people valued, but in whether they believed technology helped or hurt — and those beliefs tracked with each group’s lived experience of the devices.

Students, educators, and parents felt phones were helpful to being safe and getting help, but they differed in their view of how phones impact other values (agency and connection). Numbers represent the average dyad score from +50 (“Increases stress and anxiety to 10/10”) to –50 (“Decreases stress and anxiety to zero”), for each group (students, educators, and parents), for stories strongly associated with the values of agency, connection, and safety, respectively. (n= 874)

Parents and teachers strongly associated phones with stress. Youth were mixed: some told stories of phones adding stress, like a middle schooler whose sleepover unraveled when friends pulled out their phones. Others retold stress-relieving moments, like accessing reading supports or connecting with overseas relatives.

Strikingly, students were far more likely than adults to tell stories of using technology to meet someone else’s needs. Against the stereotype of the self-absorbed “screenager,” they used their phones to care for others — sometimes in everyday moments, and sometimes in a crisis, like in this story about a peer’s mental health crisis:

“ Me and a few other kids used our cell phones, while in class, to coordinate how we were gonna bring this to the counselors. In the end, she got the help that she needed.”

Students were more likely than adults to tell stories about helping other people. Color contours show the “hot spots” for each group. The more toward the red end of the spectrum, the more stories were clustered there. “Educators” includes teachers, district staff, and school administrators and staff. Educator n=207. Parent n=474. Student n=154.

Adults rarely told stories like this. Theirs more often reinforced familiar tropes of distraction and addiction — though the sheer volume of negative teacher stories about classroom disruption signals a real challenge leaders shouldn’t ignore. Stories like one titled “Repeat Repeat Repeat” narrated the exhaustion of teachers’ daily battles with students over devices:

“Teacher reminds students to put up phones and earbuds as they walk in. Some students do, some do not, and some lie… During class teacher sees four students with earbuds hidden under their hair. When redirected again one student pulls them out and puts them away. Three others do not.”

The enforcement struggle is real. But the near-absence of other kinds of adult stories also hints at an ingrained mental model assuming youth are self-centered and impulsive — missing their full reality.

A Shared Problem

What hit hardest were the stories adults told about their own phone use. The national narrative pathologizes student behavior, but adult accounts revealed that cell phones make adults feel out of control, too. And for parents, phone distraction can trigger parental guilt, like one parent wrote in “Mornings at home”:

“My kids are telling a story and one of them asks me ‘mommy, do you need to be on your phone?’ I had gotten a text and was paying more attention to my text than to my kid’s story.”

What emerges from the data is not a community at war over cell phones. It is a community of people who care about the same things, struggling in different ways to live well alongside a complicated technology.

A teacher worn down by daily enforcement and a parent thrown off by their own screen use can reach the same verdict — take the phones away — for very different reasons. Some of that adult urgency is a need for boundaries we feel in ourselves and project onto our kids. The youth resisting, meanwhile, aren’t just chasing “low-nutrient” dopamine hits, as Haidt argues, but protecting a core capacity to care for others.

No one group is a neutral authority with the “right” answer. Experiences, motives, and the policy itself are all more tangled than the current debate allows — which is exactly why it calls for more humility in how we make the rules.

A Different Starting Point

The district study neither affirms nor refutes screen bans in schools; instead, it changes where the policy conversation should begin.

The data challenges the paradigm of young people as passive recipients of adult decisions about technology. Student stories revealed principled reasoning, awareness of tradeoffs, and genuine concern for their peers. Policies that account for their reality are likelier to work. Enforcement friction isn’t inevitable; it’s a design problem, and the fix is to bring students into the design.

For district leaders: start conversations that surface lived experiences, not just pre-formed opinions; create policy advisory groups spanning the full range of perspectives; build from common values; and model communication that treats the challenge as a shared one, not a youth deficit.

For state policymakers, address the blind spot. Restrictive technology policies aren’t necessarily wrong. But those made entirely from adults’ point of view (without authentic youth input) and imposed without implementation support produce the push-back and unfulfilled promises already in the data. Investing in community engagement and real co-creation with students would yield more durable outcomes than a top-down rule.

The youth mental health crisis is real, and so are the concerns driving policies shaping how students use phones, screens, and AI in schools. What’s missing — and what matters most to getting policy right — is what youth themselves know about their lives with technology.

It is worth asking them.

About the report:

This blog post and the report’s findings represent independent research by CIE and do not imply official district positions, policy advocacy, or endorsements by the district or any other entity. Read the full findings and learn more about our methods here.

So over the past two months, I’ve prioritized sitting one-on-one with dozens of teachers as the schools I support closed out this academic year and began turning their attention to the new year.

In these empathy interviews, I also asked them to share their concerns and the pinch points that kept them or their students from greater success.

To be honest, this season of listening was hard for me. The depth of despair and disconnection was more severe than I have experienced since I began working in schools in 2000. Never have I seen educators so willing to distance themselves from the students who struggled most.

As a listener, my role was not to refute their analyses but instead uncover what sat beneath their frustrations. I would use probing questions to explore their experiences, motivations, personal beliefs, and emotions to give more depth to their requests for more punitive school culture practices.

I’d ask probing questions like:

Can you share a time when a student struggled to meet expectations? How did that feel? How would you describe your relationship with that student? What do you think is going on for that student? What are your hopes for your classroom next year? What do you hope to be true for this community in five years?

When I listened to understand the depths of their frustration, teachers consistently moved beyond their essentialist perspectives, simplified solutions, and flattened analyses and towards a much more complex accounting of the dynamics at play. They were far more forthcoming about their feelings of despair, inefficacy, and isolation than when the conversation first began. Where they had placed agency and responsibility for change on administration and students, they began to embrace that the solutions to student engagement lay in shared responsibility — including their own agency to strengthen relationships and create the conditions for a positive classroom climate.

They were able to hold multiple truths at the same time: students benefit from clarity and consistency in expectations and accountability, and student engagement is deeply tied to their relationships with the teacher, peers, and schooling. We moved from discourses of deficits, blame, exclusion, vilification, and “one right way” towards a discourse that embraced community assets, shared responsibility, shared leadership, community, and complexity.

We didn’t move past the discomfort. We moved through it. That made all the difference.

Again and again, I experienced educators moving beyond the flattened stories they arrived with. Not because I convinced them they were wrong. Not because I introduced a better framework. Simply because we stayed with the conversation long enough to uncover what lay beneath their initial analysis. The story underneath was almost always more human than the story they began with.

The Open Futures Initiative exists for exactly this reason.

This spring Landon Mascarenaz and I facilitated an Open Futures Simulation with leaders from the Leadership Institute of Nevada. On the surface, the session was anchored on a simulation of a district navigating declining enrollment, budget reductions, and the possibility of school consolidation. Leaders stepped into different roles, wrestled with competing priorities, and worked toward solutions that no single stakeholder could have designed alone.

The simulation created the conditions for exactly the kind of social friction Doannie Tran describes in his recent post. Leaders had to wrestle with competing perspectives rather than rushing toward agreement. They had to practice empathy across difference. They had to work toward solutions that no single stakeholder could have designed alone.

But before anyone entered the simulation, Landon invited leaders into a different kind of conversation. We shared stories about the people who shaped us, reflected on why we chose this work, and sat with a question that framed the rest of the day:

Can we build an open democracy without open hearts?

As the day unfolded, Landon brought us back to this larger frame. Inviting leaders to reflect on their experiences, their journeys, and the challenges they face. They spoke honestly about the moments when they struggled to remain open-hearted. They wondered aloud whether they were making the difference they hoped to make. They named the despair that comes from watching systems continue to underserve young people. There was also laughter, joy, humor, and hope.

We didn’t treat these emotions as distractions from leadership. We treated them as essential to it. The day made room for gravity without becoming heavy, and for joy without becoming superficial. Leaders didn’t have to choose between vulnerability and problem solving. They could hold both.

We rush past the conversations that help people make meaning together.

Looking back, I don’t think we simply designed a better simulation. We created the conditions that allowed leaders to stay with the questions they are usually asked to move past. Questions about purpose. Questions about despair. Questions about hope. Questions about what it means to remain open-hearted while leading through extraordinary times. Those conversations didn’t compete with the technical work of leadership. They gave people a deeper reason to do the technical work together.

The dominant discourse of leadership rewards urgency, efficiency, and implementation — moving us quickly toward solutions and away from the very conversations that help people make meaning together. We rush past story. We rush past emotion. We rush past uncertainty because it doesn’t fit neatly into an implementation plan.

Reading Doannie’s recent posts helped me realize that if social friction helps us stay with one another, there may be another dimension of productive friction that deserves our attention. It is the friction of staying with ourselves long enough to move beneath our first reactions, our simplified analyses, and our flattened stories. It is the willingness to linger with grief before turning it into blame, with uncertainty before turning it into an implementation plan, and with audacious hope before turning it into strategy.

Not pitting technical work against human work. Holding them together.

We don’t build better systems by setting aside the human work. And we don’t do the human work by stepping away from the need to improve our systems. Creating conditions for people to stay with their emotions, complexity, and personal truths allows communities to imagine and build futures they could not have reached alone.

Michael De Sousa is the Director of the Open Futures Initiative, a joint project of The Open Systems Institute and the Center for Innovation in Education. If this resonates and you want to explore what it would look like to bring this work to your system, reach out.

One of the themes running through our recent work is the cost of exclusion of critical voices in decision making and what gets decided, normalized, or lost when the people most affected aren’t genuinely in the conversation. This piece, originally published by our friends at Getting Smart, makes the same argument from a different angle: what happens when AI norms form in schools without young people ever being asked to help shape them.

In our recent posts, we wrote about good vs bad friction and human vs team-in-the-loop, so we wanted to share this again. The researchers and young people that wrote this piece share a framework for sensing, reflecting, and nudging toward healthier norms, which are emblematic of exactly the kind of co-creation we believe districts are capable of and should practice more.

This year, Ms. Chen started using AI to provide feedback on student writing. It saved her two hours a week and ensured that the last essay in the stack received the same quality of feedback as the first. She vetted every word, but she never told her students. There was no school policy requiring disclosure, and truthfully, she couldn’t quite define where her voice ended and the algorithm’s began.

When Priya received the latest comments on her essay, something felt off. The feedback lacked Ms. Chen’s characteristic blend of honesty and warmth that had always left Priya feeling seen.The next time around, Priya did not submit her draft early for feedback. She never brought it up, but something between them shifted.

This quiet rupture is increasingly common as AI spreads into classrooms. Ms. Chen isn’t cutting corners; she’s managing an impossible workload. Priya isn’t oversensitive; she’s a developing writer for whom that human connection is part of the learning. Both realities are true, and across classrooms everywhere, the relational ground between and among students and teachers is changing.

[Mihir Sedimbi, 16-year old from Concord, NC shares his frustration with AI-generated feedback]

But rupture is not inevitable. The alternate path begins with naming what’s happening.

Naming the Problem

Consider an analogy shared by Fitz, a 21-year-old college student. Among his friends, Fitz has become the designated “planner,” known for spring break itineraries that never disappoint. This year, he used AI to surface hidden-gem spots and weave in personalized touches. His friends declared it as his best one yet.

While battling it out in their first ever archery dodgeball game, a friend asked, “How did you come up with this? Did you use ChatGPT?” Fitz immediately said no. It was a reflex, not a decision. He then spent the rest of the week trying to figure out the rationale behind this split-second response, especially for someone who prides himself on honesty.

What he landed on was this: “AI is like Ozempic.” Not the drug, but the conversation (or lack thereof) around it. When someone announces a new diet or a gym routine, we call it discipline. When someone uses a GLP-1 to achieve the same outcome, the response is categorically different: more complicated, more loaded, carrying an implicit question about what really counts as effort. When a tool dramatically simplifies a process, the validity of the outcome is questioned.

That is the grey zone Fitz found himself in. He had used AI to plan a better trip for his friends, and his instinct was to hide it. Not because he thought he had done anything wrong, but because the norms for whether it is socially acceptable to use AI for trip planning with friend groups simply do not exist yet.

[Compare Fitz’s instinct to hide his AI-assisted planning to Alex Matthew, 17, attending Alpha school where AI is part of everyday conversation and using it to plan a spring break trip with friends felt completely natural.]

Unlike the Ozempic debate, however, the goal isn’t to take a side on AI. It is to bring young people into the conversation about how it gets used so they are not left to navigate it alone in the silence.

Why Silence is Costly

Left long enough, that silence becomes its own kind of norm. And for adolescents, the stakes are higher. Their identity is still under construction, assembled through social comparison and peer feedback. Young people are acutely attuned to how their choices are perceived and deeply invested in what those choices signal about who they are. A norm vacuum in that context produces persistent anxiety that comes from never quite knowing whether what you just did was acceptable, impressive, or something to hide.

For young people, AI is not a tool, it’s a social disruptor. It is also, notably, not the first technology to arrive this way.

As Babak Mostaghimi notes, we’ve been here before. In 2012, young people faced a different normative crisis with the mass adoption of smartphones and social media. Adults prioritized bans over guidance, and platform developers filled the norms vacuum with social behaviors designed to boost engagement (endless doomscrolling, signaling kinship with “likes,” expecting immediate responses to direct messages). Developers financially benefitted and youth mental health continues to pay the price.

Today, similarly harmful norms are now taking shape around the use of generative AI, occurring largely outside adult awareness. Some educators and organizations have already recognized this. The Harvard Center for Digital Thriving has developed classroom resources to help youth ground in their values and navigate grey areas. The Rithm Project created a card game that people are using to spark conversations about AI and human connection. These efforts point toward the importance of creating space for young people to make sense of AI on their own terms.

But for a culture shift to take hold, new norms must spread peer to peer through the social fabric of young people’s actual lives.

What a Different Approach Looks Like

Successfully shifting youth culture requires a fundamental shift in orientation that most institutions are not currently set up for. Too often, we treat AI adoption as a technical challenge, defaulting to AI literacy frameworks and handbooks.

But shifting culture is a complex, relational challenge that evolves by sensing what is actually happening, reflecting on what it means, and nudging from the inside out.

Sensing: knowing what is actually happening

The first practical step is building the infrastructure to hear from young people routinely, honestly, and at scale. Hearing from young people and actually hearing them, however, are not the same thing. Surveys can tell you what is happening, but they cannot tell you why, or what it means to the person experiencing it.

Tools like SenseMaker address this directly by asking young people to share real anecdotes about their AI experience (a time they used it, avoided it, or felt conflicted about it). Follow-up questions ask them to interpret that experience (what it meant, how it felt, whether it was typical). What emerges is a clearer picture of the harder to reach social norms operating beneath the data (Box 2). CIE is already using SenseMaker to study culture shifts in contexts ranging from cell phones in schools to workforce development. Its application to surfacing below-the-surface norms in the real-time context of AI adoption is a natural extension.

Reflecting: making meaning together

Sensing is only valuable if it leads somewhere. The second step is sharing what was heard, naming the patterns, and asking young people to weigh in on what it means and what it calls for. In practice, an intergenerational team reviews SenseMaker patterns twice a year, identifying what is moving in a healthy direction (increasing openness about AI use) and what is moving in a harmful one (concealment becoming the default). They name it in their own language, propose what to amplify and what to interrupt, and help design the nudge.

This is not a focus group or a rubber-stamp committee. It is genuine co-construction, which requires adults to show up with humility and a willingness to be surprised.

Nudging: shifting the culture from the inside

Sensing and reflecting create the conditions for the third and final step: the nudge. A nudge is a small, well-timed intervention designed to shift a norm in a specific direction.

Nudges resonate more when the people being nudged help build them. Consider the trajectory of youth anti-smoking campaigns. For years, they relied on scare tactics and grim PSAs that fell flat with teens who valued immediate social currency over long-term health. In 1998, Florida’s Truth Campaign abandoned the lecture and asked teens directly why they smoke. They uncovered an important insight: smoking wasn’t a result of ignorance, but a deliberate way to satisfy a developmental need for autonomy. By reframing Big Tobacco as the manipulative authority, the campaign flipped the script and turned the refusal of a cigarette into a defiant act of rebellion. Teen smoking dropped from 28% to less than 6%.

This same logic applies to the social currency of AI disclosure. Right now, admitting AI use carries the social risk of being seen as less capable, less creative, or less genuine. But we can move the needle by engaging young people directly. A well-designed nudge reframes disclosure from a moment of “getting caught” into an act of ownership.

Now reimagine Ms. Chen’s classroom after a few cycles of sensing and reflecting. They arrive at a core truth: submitting work is an act of trust in a specific person to provide feedback that honors the writer. They co-author a nudge: a new guideline that essays will be peer-reviewed first to ensure each one has a human audience, then given AI-assisted feedback from Ms. Chen (openly acknowledged and carefully curated). They continue to sense, reflect, and nudge as the new norm plays out to ensure it feels beneficial to all.

This agreement will not appear in any handbook. New norms cannot be mandated or scaled from above. As with any shift in culture, it ripples outward as those who experience it firsthand engage in collective meaning-making and choose to carry it forward.

About the Authors:

This blog was coauthored by authors Shereen El Mallah and Jenny Poon, and youth authors Alex Matthew and Adora Olise. Youth contributors included Mihir Sedimbi and Fitz Awiakta.

AI disclosure:

The authors wrote an initial blog outline and asked Gemini to act as an editor, tightening our arguments while retaining the overall flow and main points of the original. Working from the revised outline, the authors drafted the post without AI (examples, emdashes and all), then asked Gemini and/or Claude to copy edit, improve transitions, sharpen clarity, and reduce word count.

Imagine a curriculum director at a mid-sized district. She has six weeks to select a new literacy curriculum before the school year begins, a skeptical school board, and a stack of vendor proposals. She turns to an AI tool for help, feeds in the proposals, and asks for a ranked recommendation. The AI returns a confident, well-organized analysis with a clear top choice.

She reviews it. It’s persuasive. She’s pressed for time, and the AI’s apparent thoroughness is reassuring. She moves forward. The board approves it.

This is human-in-the-loop working exactly as designed. And it’s probably not enough.

Why individual judgment isn’t a sufficient check on AI

Human-in-the-loop has become a really common answer to concerns about AI quality, hallucination, and bias. The idea makes sense on its face: Keep a person in the decision chain, and you can preserve the critical reasoning that AI lacks. But I’ve been reading two threads of emerging research that make me think that this framing is more fragile than it appears.

At CIE we have been learning a lot from our friends at the Stanford Cognitive Security Task Force about cognitive security, the integrity of an individual’s reasoning from outside influence. According to this frame, when a person scrolls social media, their sense-making is being shaped in ways they often can’t detect. Framing, sequencing, and subtle cues can influence the meaning a person makes of what they’re reading, sometimes without their conscious awareness. Of course, this assumes that our thoughts were ever completely our own, but even if you think we’ve always been influenced by external influences, you might accept that we’re reaching a step change with AI. As AI-generated content becomes more sophisticated, these influences will only intensify, from overt manipulation like deepfakes to subtler effects that shape how people reason about what they’re seeing.

The second is cognitive surrender. Building on Kahneman’s research on System 1 (fast, intuitive) and System 2 (deliberate, analytical) thinking, researchers Shaw and Nave at Wharton have identified what they call System 3: the external reasoning now embodied in AI. Their research examines when people skip System 2 entirely and move directly from intuition to AI output without engaging their own deliberate reasoning at all. Under conditions of cognitive surrender, a person’s accuracy becomes almost entirely dependent on the AI’s accuracy. When the AI is right, they’re right. When the AI is wrong, they follow it down.

What makes this dangerous isn’t that people are careless. It’s that cognitive surrender is condition-dependent. Time pressure, incentive structures, and feedback loops all affect whether a person can maintain System 2 engagement while using AI. A curriculum director under a six-week deadline, looking at a well-organized AI output, is operating in almost exactly the conditions that research suggests produce cognitive surrender.

Taken together, these two dynamics paint a troubling picture. An individual person’s reasoning can be compromised on two fronts simultaneously: Their capacity to critically evaluate AI output, and their ability to even engage that critical capacity in the first place. I think this is going to be a consistent failure of human-in-the-loop systems: a quiet cascade of very confident, individually vetted errors.

A different model: team-in-the-loop

There’s a scene in Guardians of the Galaxy where a villain seizes a powerful stone (the Power Stone for the MCU faithful) and it destroys him, because no single person can handle that much power alone. But when the Guardians link arms and channel it collectively, they can contain the power relatively safely. As both a nerd and former science teacher, I live in metaphors, and I think this is a useful image for what we’re proposing as a meaningful alternative to human-in-the-loop.

Human-in-the-loop, as currently conceived, asks one person to wield the stone. What we need instead is a team.

Not a team as a formality, but a team as a structured check on the conditions that produce cognitive surrender and a check against cognitive intrusion. What a team actually brings is diversity: of expertise, of perspective, and crucially, of skepticism. Research consistently shows that diverse teams produce higher-quality outputs and more creative solutions. What we’re arguing is that a team-in-the-loop also produces something else: a distributed resistance to cognitive surrender, where varying levels of skepticism and different knowledge bases can pressure-test AI output in ways no individual can.

Return to the curriculum director. Under a team-in-the-loop model, she brings the AI’s analysis to a working session: her literacy coach, a veteran third-grade teacher, a parent from the district’s equity committee, a special education coordinator.

The AI’s top-ranked curriculum scores well on cost and state standards alignment. But the literacy coach noticed the AI weighted vendor-supplied efficacy data heavily and discounted a study showing mixed results for English language learners. The special education coordinator flags that the curriculum’s pacing assumptions are incompatible with IEP flexibility. The parent representative asks why cultural responsiveness wasn’t a criterion at all.

None of these people would have caught everything alone. Together, they caught what the AI couldn’t see, not because the AI was wrong exactly, but because the questions it knew how to answer weren’t the only questions that mattered.

What this requires in practice

At CIE, we have been experimenting with what it looks like to put a team-in-the-loop process into play. In our internal team processes and work with client organizations, we’ve been trying to understand what it takes to engage with potential biases and keep our System 2 deliberation online. We’ve discovered some things we want to share. Team-in-the-loop isn’t just about who’s in the room. It requires deliberate design and facilitation. A few things matter most:

Protect the conditions for System 2 thinking. Time pressure is one of the most reliable triggers of cognitive surrender. Leaders who want genuine critical engagement from their teams need to create containers such as structured working sessions, professional learning communities, review protocols that shield the team from the pressures most likely to short-circuit deliberation.

Feed learning back into the system. A team-in-the-loop process generates something valuable: a record of where AI got it wrong, what it missed, and what the team caught. That learning should be systematically captured and used to calibrate how AI is used going forward. Those who are facile with AI can cycle these insights into guidance that can inform the model, but they can also just be guidelines for human users.

Build psychological safety. None of this works if team members don’t feel safe surfacing concerns. People need to be able to question assumptions and assertions without the fear of being seen as “troublesome” or “slowing things down” or “preoccupied with process” (of course this requires that people don’t try to overly manipulate things through process, but that’s an issue for different post). Oftentimes, the person most likely to catch a consequential AI error is also often the person with the least institutional power. Designing for these realities isn’t a soft culture issue, it’s about whether your system is organized for learning.

Examine your incentive structures. A relentless focus on speed and quantity of output, applied uniformly across all decisions, is a perverse incentive for cognitive surrender. Not every decision warrants the same depth of scrutiny, but high-stakes ones such as curriculum adoption, policy creation, strategic planning, need explicit protection from the kinds of pressure that lead people to just go with the AI’s answer.

Which actually leads to the last thing:

You can’t put a team-in-the-loop for everything. Deliberation is difficult and slow and most teams aren’t currently organized to apply that kind of scrutiny to more than a limited number of critical projects. Prioritizing is actually a really powerful task in and of itself - a place to name common agreements about how AI will be used within the team.

The harder question

The curriculum director who moved forward with the AI’s recommendation wasn’t negligent. She was busy, under pressure, and working with a tool that projected confidence. The human-in-the-loop model gave her what she needed to satisfy the request she faced.

If the implementation failed, it’s not because she failed, but because the model asked too much of any single person and didn’t figure out how best to leverage the team around her.

The leaders who navigate this well won’t be the ones who distrust AI. They’ll be the ones who understand what individual human judgment can and can’t do in relationship to it. We think that strong leaders will build the structures that let teams catch what individuals miss - if they can acknowledge that’s not a technology problem, it’s a leadership one.

I was recently talking to a respected colleague about using AI to support community engagement in strategic planning. Reacting to an AI tool CIE is building, she lit up and suggested we could use it build digital twins representing typical community members and run stakeholder interviews with them, saving the time and messiness of talking to actual people. It was a beautifully modern idea: efficient, safe, convincing. And precisely the wrong move.

We invest time and money teaching educators how to cultivate “productive struggle” in classrooms, guiding young people to know when AI is helpful and when true learning requires old-fashioned cognitive discomfort. But that moment showed me that we haven’t made the same investments in school system leadership. When it comes to community engagement, we risk outsourcing the kind of struggle that matters most.

This is the defining leadership challenge of the AI era. Not whether to use AI, but whether leaders can tell the difference between friction that slows you down and friction that holds everything together. Leaders should eliminate the former, but they need new tools to fashion the latter into something worthwhile.

We’ve been removing the wrong friction for decades.

Long before AI arrived, districts developed a comfortable set of procedures. Strategic planning processes tended to generate the appearance of community voice without the substance of it. Surveys helped districts make claims about what families want, allowing leaders to explain findings through the lens of their own priorities. Administrators built dashboards to report to other administrators more so than to the community.

These practices share a common assumption: that the hard interpretive work of making sense of what a community actually needs can be delegated to an easy to use tool or a closed-door meeting. This assumes that the friction of real collaboration, with all its disagreement, ambiguity, and the slow work of building trust across differences, is a barrier to be engineered away rather than a resource to be cultivated.

Districts have conflated token “engagement” with co-creation for so long that they’ve stopped noticing the gap. But every time a decision lands without authentic buy-in, or a solution turns out to miss what the community actually needed, or an initiative stalls because nobody feels ownership over it, the district experiences the cost of that shortcut.

AI doesn’t create this problem. It accelerates it. It makes frictionless automation faster, cheaper, and more convincing. Which is why the leaders who get it right won’t be the ones who accelerate everything. In fact, they’ll be the ones who know what not to accelerate.

Good Friction vs. Bad

Friction is what happens when something is hard, sticky, uncertain, or contested. The current energy in AI adoption is almost entirely oriented toward eliminating it. And that makes the kind of sense that turns out to be wrong, like an F1 car engineered to go faster by removing the brakes.

Some friction genuinely gets in the way: administrative burden, logistical bottlenecks, the tedium of tasks that don’t require human judgment. AI should eliminate as much of that as possible.

But some friction is load-bearing. The friction of sitting across from someone whose experience differs from yours, and working through it. The friction of bringing together a room that won’t agree. The friction of a question that doesn’t have a clean answer yet. This kind of friction is tied to trust and relationships. Deliberation shouldn’t be considered a bug in the process. It’s where the real work happens. It’s how communities construct something they actually own.

The most frictionless engagement strategy is to let community members remain in their separate corners, offering up isolated opinions for leaders to heed or ignore based on their own judgment. But that is a shortcut that generates enormous risk. When leaders or institutions do this, they bear all the consequences alone. The frictionless decision begets greater resistance during implementation.

When diverse perspectives are deeply engaged, however, they work through friction on the front end, bringing a different kind of legitimacy that distributes the risk, builds trust, and increases shared ownership. The tolls paid up front through good friction pays off in acceleration when it matters - in implementation.

The leaders who will be most trusted in the AI era aren’t the ones who eliminate all friction. They’re the ones who manage it and use AI to reduce the friction that wastes people’s time, while preserving and managing the friction that builds understanding and commitment. What we need is a new generation of leadership tools that don’t just engineer away the messiness, but help leaders actively manage the friction that makes a decision durable.

What this looks like in practice.

In Vicksburg Warren School District, we used four habits of inclusion, empathy, co-creation, and reciprocity to build an accountability dashboard with the community, not for it.

We didn’t create a digital twin of a coalition. We put together a diverse team and used deliberate processes to manage the friction of working across different perspectives to find consensus. We did the hard collaborative work of identifying metrics that actually mattered to the community. I did use AI to help develop a working prototype quickly. That was good friction management: using AI to reduce the unhelpful kind.

But here’s the moment that mattered.

When we had a working prototype, it would have been easy to ask the people in the room: Does this match the metrics we identified? That question would have gotten a solid yes. It would have produced a satisfied client and a finished product - friction free.

Instead, we asked: Does this address the things you actually talk about in hallways, in parking lots, at the grocery store?

That question invited friction. It opened space for uncertainty, for honest reaction, for things that didn’t fit the template.

A student told us that school got boring somewhere between elementary and middle school when the work stopped feeling challenging. A parent talked about the feeling of disorder that lives beneath “safety” numbers, something that was felt but never measured.

None of that would have surfaced from a digital twin. None of it would have come from a feedback process designed to produce confirmation. Moreover, even if it had, without that room full of divergent perspectives, no one would have heard it, seen the district respond to it, or built the kind of shared product that inspires confidence.

What we built because of those conversations was different: a dashboard that includes data from student advisory councils, staff task audits measuring relevance and rigor, and school-based empathy interviews gauging students’ sense of order and safety. Moreover, building it together, friction and all, created the energy and momentum to ensure that it gets used. Friction does generate heat, after all.

The question for leaders.

AI will keep getting better at producing things that look like community engagement, and will do so more quickly and with an astounding amount of asserted confidence. The pressure to substitute efficiency for substance will only increase.

The leaders who build genuine trust and make decisions that actually hold will be the ones who can answer a harder question: Where in our work are we removing friction that we actually need, and how might we engage our community in a productive struggle to create something worthwhile?

That’s not an argument against AI. It’s an argument for using it with a clearer theory of what it’s for.

The friction that matters is the kind that builds trust, develops understanding, and leads to decisions communities will actually stand behind. It isn’t a problem to be solved. Friction is what allows you to turn a corner, to move with purpose.