The era of schools buying every shiny new AI tool is ending. After a decade of accumulating thousands of educational technology products with little evidence they improved student outcomes, districts are now asking harder questions about what stays and what goes.

Why Are Schools Rethinking Their AI Adoption Strategy?

Parents and educators are pushing back. They want to know whether the tools in classrooms actually help students learn or simply add distraction, cost, and complexity to the school day. The shift reflects a broader recognition that technology adoption should not be the goal; improved learning should be.

This reckoning comes as AI takes center stage in education. More than 60% of middle and high school teachers in core subjects like English, math, and science are using AI at school, according to a 2025 report from the RAND Corporation, a research organization. Elementary school teachers lag behind, with only 42% ever introducing AI in the classroom.

The problem is that little research exists on the best ways to teach young children about AI and which skills should be prioritized. Schools are moving forward without a clear roadmap, and experts are divided on the right approach.

What Five Questions Should Guide School Technology Decisions?

As district and school leaders revisit technology policies and respond to growing pressure from parents, educators, and policymakers, they have an opportunity to replace reaction with rigor. Five key questions should guide their decisions:

• Start with a Problem, Not a Product: Technology should not be a given in classrooms. It should earn its place by demonstrating that it can improve teaching and learning. Too often, technology decisions begin with a vendor pitch or pressure to innovate. Instead, districts should ask: Is the goal to help kids who are behind in math? Provide faster feedback on student writing? Expand access to tutoring? A tool should only be considered if it is tied to a clearly defined challenge and aligns with the district's goals.
• Demand Credible Evidence: Not all screen time is equal. Districts should review testimonials and usage data, but they should not stop there. It is important to understand whether a tool has demonstrated impact on student outcomes, in what context, and for which children. This evidence should initially come through small-scale studies and then build up to more rigorous third-party research that has been vetted externally.
• Ensure Technology Supports Teachers, Not Replaces Them: Human relationships are central to how students learn and are likely to become even more valuable in an AI-enabled world. No technology should replace the role of a skilled teacher, the value of peer debate, or the importance of productive struggle. If a tool weakens the relationship between teachers and students, districts should be skeptical. If it helps educators better understand student needs and free up time for interaction, there may be a place for it.
• Define Success Metrics Before Adoption: Districts have developed sophisticated processes for procuring technology but need equally rigorous processes for removing tools that fail to improve outcomes. Before adopting a tool, leaders should define what success looks like, over what time period, and how they will know whether the tool is helping. Schools should be willing to walk away from tools that do not demonstrate value in the expected time frame.
• Communicate Clearly with Families: Families deserve to know which technologies are being used, why they were selected, what evidence justifies them, and how they are supporting their child's progress. This is especially important for AI-enabled tools, as parents may reasonably worry about privacy, academic integrity, and overreliance. A district should be able to explain in plain language how a tool accelerates learning.

"The next phase of educational technology should be defined less by adoption and more by how carefully districts choose what stays," according to analysis from The 74, a nonprofit news organization focused on education.

How Are Schools Teaching AI Literacy to Young Students?

One emerging approach is teaching AI concepts without heavy screen time. Overture Games, a startup founded by two recent Northwestern University graduates, is working with 36 schools and youth centers across Illinois, plus another 18 in Massachusetts, using a paper-and-pencil method to introduce younger children to AI skills.

The approach works like this: students brainstorm ideas by hand, write prompts on worksheets, and then instructors input their notes into AI tools like Suno (a music composition tool) or Google Gemini. Students see the AI-generated output and learn to refine their prompts. By limiting direct student access to AI systems, the company argues it creates a safer learning environment while still exposing kids to the technology.

"As soon as they hear that kids are not directly introduced to AI tools themselves, meaning putting their hands on it, just doing whatever they want with it, that is the definition of safety to schools and parents," said Steven Jiang, cofounder of Overture Games.

Steven Jiang, Cofounder, Overture Games

In practice, students in Overture's after-school programs learn to spot common AI mistakes. They play games like "Real or AI" to distinguish between photos and AI-generated images, noticing details like how AI struggles to render realistic body parts or often adds excessive lighting. One 8-year-old participant, Matthew Uriosdegui, even improved his spelling by 25% after learning that AI tools produce lower-quality outputs when prompts contain misspellings.

Do Experts Agree on How to Teach AI to Young Children?

Not entirely. Victor Lee, an education professor at Stanford University who helps high school teachers incorporate AI skills into their lesson plans, questions whether prompting is the most important skill for young students to learn. He believes elementary schools should teach kids to use AI only if "there was a really valid and productive instructional purpose tied to important learning goals".

"The question we have to ask is: Is prompting the most important skill to learn? I would say a lot of AI literacy experts say no," explained Victor Lee, education professor at Stanford University.

Victor Lee, Education Professor, Stanford University

Lee suggests that conversations about common AI mistakes are more valuable than encouraging young kids to actually use AI tools. However, Elizabeth Radday of EdAdvance, a nonprofit that supports public schools in Connecticut, argues that a hands-on approach helps students retain information better than passive learning.

"There's a big difference between kids that are doing things to create a final project, versus clicking through something and just watching videos, or learning about AI through watching someone else do it," noted Elizabeth Radday, EdAdvance.

Elizabeth Radday, EdAdvance

The disagreement reflects a broader uncertainty in the field. Little research exists to guide schools on how best to teach AI literacy to young students, and access to these programs remains uneven across districts.

What's Happening at Major Education Conferences?

The conversation about AI in schools is taking center stage at major education events. ISTELive 26, a conference for K-12 instructional staff, technology leaders, superintendents, and librarians, will be held June 28 to July 1 in Orlando, Florida. The conference will feature dozens of AI-focused sessions and an interactive "Profile of an AI-Ready Graduate" exhibit where educators can see demonstrations of AI-enabled learning in action.

Two sessions in particular will explore how schools are developing the skills students will need to collaborate with AI after graduation. "Tomorrow's Creators Today: Creating a Future-Ready AI Graduate" and "Profile of an AI-Ready Graduate: Durable Skills for Learning and the Workforce" will emphasize why building critical-thinking skills will be vital for students to collaborate with AI now and into the future.

Beyond AI, the conference will address broader technology trends, including how districts can secure funding for technology upgrades, balance edtech policy with innovation, protect student data privacy, and teach media literacy and critical thinking in an age of AI.

The backlash against technology in schools has identified real harms. Parents are right to be concerned about tools that undermine attention, replace human interaction, or encourage passive use. But the lesson of the past decade is not that technology is bad. It is that schools need to be far more disciplined about what they adopt, why they adopt it, how they measure success, and how they communicate those decisions to families.