From Fringe to Policy: How AI Is Shifting the Overton Window in K-12 Education

When you hear the same unfamiliar term twice in one week you start to wonder if something’s shifting in your field.

During an interview for an upcoming magazine article, a tech executive told me that he was writing a white paper describing how policy ideas in AI implementation could be explained by using an “Overton Window.” Three days later, the New York Times published a piece that detailed how the Overton Window can be used to understand the radical changes being proposed to K-12 ed policy. Hmm. 

The Overton Window is a protocol used to categorize ideas or policies that are considered acceptable to talk about in public at any given time. Imagine it like a sliding window on a wall of possibilities — only the ideas inside the window are seen as “mainstream” or realistic, while those outside it might seem too radical, outdated, or extreme. What’s interesting is that the window can shift over time — what once seemed unthinkable can become normal if enough people start discussing it or if the context changes. 

The Mackinac Center Tees Up the Conversation

Let’s begin with polarities being discussed by the Mackinac Center through the lens of the Overton Window. The Center frames the education debate using intentionally extreme endpoints: From “zero public investment in education” to “compulsory indoctrination in government schools.” These rhetorical bookends aren’t policy proposals — they’re conceptual poles meant to define the boundaries of the debate.

The process requires us to imagine all possible education policies ordered from minimal to maximal government involvement, then note that only some segment of that spectrum is politically possible at any time.​ Joseph P. Overton, a Mackinac Center executive in the 1990s, used this model to explain to donors that think tanks cannot simply “pick” any policy they like; elected officials are constrained to options that fall within what the public will accept.  

Education is a natural case for this example because policy choices range from highly centralized state provision to largely private systems with minimal state involvement. The two endpoints — no government role at all versus fully compulsory, ideologically loaded state schooling — are intentionally exaggerated to make clear that the window is about theoretical limits, not actual proposals anyone expects to enact. Yet. 

Here is another example, near to my heart, that illustrates the movement from fringe to mainstream in education.

At the turn of the 20th century, most schooling was rigid, teacher-directed, and focused on rote memorization. Reformers like Francis Wayland Parker and John Dewey pushed a radically different idea: Schools should focus on the whole child, not just the transmission of facts, with learning rooted in real experience, problem solving, and cooperation rather than memorizing textbooks. At the time, these ideas were dismissed by many as too soft or impractical. Over decades, however, elements of this progressive philosophy — hands-on projects, inquiry-based learning, and collaborative work — became embedded in mainstream debates about curriculum and pedagogy.  

The New Fringe, Different Than the Old Fringe

When Overton developed this process he envisioned a continuum ranging from “Unthinkable” to “Policy.” Ideas that are within the window (mainstream) are those that a politician (or school leader) can recommend without appearing too extreme in the eyes of the electorate. Ideas outside the window (fringe) are those that would pose significant negative consequences or jeopardize re-election due to the idea being widely perceived as too extreme or unacceptable.  

What qualifies as a “fringe” idea in the AI era has grown increasingly fuzzy. Did you imagine 18 months ago that we would be discussing teacherless classrooms and two-hour instructional windows led by chatbots? 

I relish the chaos. A 12-year classroom career as a history teacher has nudged me toward ideas once considered heresy: Do we really need to teach children under the age of 10 names and dates from the ancient past? Their temporal context window is less than a decade, yet we expect them to comprehend the importance of events that occurred centuries ago. Nonsense. 

I’ll stop there to answer hate mail while turning to my AI thought partners to come up with a list of the unthinkable policies we can insert into an Overton Window slider. 

In no particular order:

1. AI-Only Classrooms in Low-Population or Rural Districts

Fringe Idea: Eliminate in-person teaching in low-enrollment areas and replace it entirely with AI instructors and AI tutors, overseen by a single adult “learning facilitator.”

Why It’s Fringe: It directly challenges the assumption that all students need a certified, human teacher for daily instruction.

Path to Mainstream: Chronic teacher shortages, shrinking rural populations, and budget crises could open the Overton Window to this as a “last-resort” policy.

2. AI-Created, Student-Personalized Curricula That Replace State Standards

Fringe Idea: Scrap grade-level standards model and allow generative AI to dynamically create personalized learning pathways and assessments for each student, based on their interests, pace, and performance.

Why It’s Fringe: State standards are deeply embedded in educational policy, accountability, and funding. This would blow up the current system.

Path to Mainstream: If AI systems show higher learning gains through personalization—and policy fatigue around testing continues—this could gain traction.

3. AI-Powered Learning Pods Replacing Traditional Schools

Fringe Idea: Families opt out of public or private schools and join AI-supported micro-learning pods, where instruction is delivered by AI and supervised by community members or paraprofessionals.

Why It’s Fringe: It undermines the very structure of public schooling, teacher certification, and school funding models.

Path to Mainstream: Post-pandemic pod models, combined with declining trust in institutions and rising AI capabilities, could shift this into serious debate territory — especially in choice-heavy states.

4. AI-Generated Emotional and Behavioral Support Replacing Traditional Counseling

Fringe Idea: Replace or supplement school counselors with AI-driven emotional support bots trained in CBT, trauma-informed practices, and local cultural norms.

Why It’s Fringe: Deep skepticism remains about AI’s ability to address human emotional complexity, especially with vulnerable populations.

Path to Mainstream: If demand for mental health services continues to outpace supply and AI empathy tools improve, this could become a lower-cost solution.

5. AI-Powered Teacher Hiring and Evaluation

Fringe Idea: Use generative AI and machine learning to screen teacher applications, analyze classroom recordings, and provide evaluative feedback that informs tenure decisions, professional development, or even layoffs.

Why It’s Fringe: This raises significant ethical, labor, and privacy concerns. Many see it as dehumanizing or hostile to educators.

Path to Mainstream: If districts adopt AI-based HR systems for hiring and observation — as some are quietly piloting — the policy could shift from outrageous to inevitable.

There are a handful of forces that can move these ideas from the fringe to mainstream, among them:

• A crisis or disruption
• Cultural shifts and generational turnover
• Technological breakthroughs
• Media amplification
• Political opportunity
• Influence of thought leaders and early adopters
• Economic incentives and market forces

Buckle up, because nearly all of these forces are in play today.

An Overton Window App

This led me to wonder: Could we use AI to bring the Overton Window to life as an interactive policy tool?

I used Gemini and ChatGPT to brainstorm and then refine my ideas. I chose Claude Artifacts, my new favorite tool for creating interactive educational tools, to code my Overton Window tool for education (see screenshot). 

How to use it:

1. Select one of the “fringe ideas” I listed in the prior section by clicking on the labeled button (“AI Learning Pods,” for example) and see where the Overton slider places it on the continuum from Unthinkable to Policy/Popular.
2. Or, type a new scenario in the text box (e.g., “Mandatory Coding Classes K-5”).
3. Click “Add Scenario” or press Enter
4. A new button appears with your scenario
5. Click the custom button to move the window to wherever it was positioned when you created it.
6. Click the × on any custom button to delete it

Each custom scenario “remembers” where the Overton Window was positioned when you created it, so clicking the button will return the window to that exact position. This allows users to explore and save different policy positions.

As you can imagine, this tool can be used as a discussion prompt for staff, parents, board, or leadership meetings. Bring popcorn. 

Final thoughts

Organizations and companies pay me a lot of money to tell them where education is going in the age of AI. That might not be the best use of resources because in a fast-shifting AI landscape, yesterday’s assumptions are today’s constraints. I argue, often successfully, that AI changes the context so rapidly that conventional thinking is no longer an effective strategy.

As of late 2025, roughly two‑thirds of U.S. states have some form of AI education policy or guidance in place for K‑12, but there is not yet a single, authoritative statistic on the exact percentage of school districts with a formal AI policy. Current evidence suggests that a minority of districts have a fully articulated AI policy, even though a majority are already using AI tools. Related to which, the Ohio legislature just mandated that all districts in their state must have AI policies in place by July of 2026. 

There are countless tools that can guide this policy discussion. Among my favorites are the AI Policy Labs, CoSN’s District Leaders Action Summit model, ISTE’s AI hub, and Digital Promise’s AI in Education initiative and AI Literacy Framework.

If your state or district is dragging its feet, I suggest the Overton Window protocol and my Overton Window interactive slider as a starting point for the conversation. As these discussions occur, it would be helpful to gain clarity around which policies and practices are within the realm of political possibility in your context and which are not.