Schools test AI, but teachers still lack use cases

A two-year study with teachers finding generative AI adoption remains minimal—because many educators can’t pin down a clear, lasting way to use it. The research points instead toward computational thinking as the skill most likely to endure as tools evolve.

For more than a decade, schools have been chasing the skills they think the “future of work” will demand—first through coding initiatives often built around the idea of “Hello, world!” and now through the latest wave of generative AI urgency.

But when teachers tried to translate that urgency into classroom practice. the story looked less like a breakthrough and more like a pause. In a two-year research project conducted alongside teachers who said they were open to integrating AI. the study found uptake was still minimal. Most participants— including teachers who described themselves as engineering or computer science teachers—struggled to identify a clear or universal instructional use case for widespread AI integration.

The lesson is not that educators are refusing to change. It’s that the central promise of AI in classrooms—an immediate, teachable, scalable method—hasn’t landed cleanly.

Schools weren’t always dealing with uncertainty. In the early push for computer science education. districts introduced new courses. nonprofits expanded access to computer science education. and a growing ecosystem of programs promised students the skills needed to enter the tech workforce. For many, it felt like the right correction for a digitizing world.

Over time, though, a more complicated picture emerged: the relationship between early coding exposure and long-term workforce outcomes became uneven. That experience helped keep alive a question that never fully disappeared— which skills actually endure when technologies change?

That question has resurfaced with generative AI. Schools are again being encouraged to adapt quickly. often using the same broad rationale: prepare students for a future shaped by emerging technologies. Yet the study’s findings suggest that. even when teachers are willing. figuring out how AI fits instructional goals is still the hardest part.

Discussions about AI education have often focused on teaching students how to use generative tools effectively. Prompt engineering—highlighted in professional development workshops and online tutorials—has become a common topic.

But the research describes a trap teachers have seen before. When curricula center too heavily on tool-specific skills, the result can be familiar: the interface changes faster than learning objectives. Teaching students how to interact with a particular AI system risks becoming “the equivalent of teaching to standardized tests. ” even if the deeper lessons don’t appear on state exams.

The computing education history the study draws on makes that warning concrete. In the early 2010s, coding initiatives encouraged schools to teach programming skills broadly. While those programs expanded access to computer science education, later analysis showed technology workforce pipelines stayed uneven. Many students learned tool-specific skills without developing deeper computational reasoning abilities.

That is the caution the research brings forward to the current AI moment: if the goal is long-term preparation for technological change, focusing narrowly on how to use today’s tools may not be durable.

Instead, the research points toward a different target—computational thinking—as a more enduring educational objective. Computational thinking is described as a set of problem-solving practices used in computer science and other analytical disciplines. including breaking complex problems into smaller components. recognizing patterns. designing step-by-step process. and evaluating the outputs of automated systems.

These skills, the study argues, apply beyond programming. They can be used in fields ranging from engineering to public policy. Just as importantly. computational thinking helps students understand how algorithmic systems operate. allowing them to analyze how technologies like AI produce results rather than simply treating those results as authoritative.

In classrooms. the teachers in the study were already moving toward that kind of thinking. sometimes without using the term “computational thinking.” When teachers asked students to analyze chatbot errors. they were pushing students to examine how algorithmic systems produce outputs. When teachers designed exercises comparing training data and algorithms to everyday processes. they were training students to reason about how automated systems work.

These approaches don’t require students to lean heavily on AI tools themselves. They treat AI as a case study—something students can analyze for what it reveals about how information is shaped.

That framing connects directly to longstanding educational goals around critical thinking, media literacy, and problem-solving. The study describes approaches such as using AI systems as objects of analysis: asking students to evaluate outputs. identify errors. and investigate how models generate responses.

It also connects AI learning to broader topics that can be taught whether the tool changes tomorrow or not—data quality. algorithmic bias. and information reliability. Assignments that emphasize reasoning. structured problem solving. and evidence evaluation are presented as ways to keep students doing the cognitive work that remains central to learning.

The study’s caution is paired with a practical boundary: these approaches allow students to engage with AI without letting the technology replace the thinking process itself.

Teachers are not the only group affected by the uncertainty. The research also places responsibility on the product side—pointing to an opportunity for edtech developers. Because many AI tools were developed as general-purpose language systems and later introduced into education. teachers are often left to determine whether and how those tools align with classroom learning goals.

In the study’s conversations, teachers were already experimenting with small classroom applications—designing AI literacy lessons and building course-specific chatbots—but the research suggests these efforts often resemble early-stage product testing.

The implication is clear in the way the study describes possible collaboration: partnerships between educators, edtech developers, and product managers could help identify instructional problems AI systems can realistically address.

The project also looks ahead. The study describes this series of conversations as an early attempt to document how teachers are navigating the arrival of generative AI. and the next challenge as developing governance frameworks—rules and guidance that help educators evaluate when and how AI should be used in learning environments.

In its next phase. the research team says it is partnering with school districts to develop guidance for AI governance and is inviting edtech companies interested in exploring these questions collaboratively. The planned focus is not on assuming AI will inevitably transform classrooms. It will instead look for the conditions under which AI tools support teaching and learning and how to reduce harm when they do not.

For teachers, the question that keeps resurfacing is simple enough to fit on a chalkboard: the study points back to “the fourth grade teacher’s question,” What can I actually use this for in math?

Until that answer becomes clearer across classrooms, the study suggests many educators will keep doing what professionals do when new technologies appear—experimenting cautiously, adopting what works, and relying on their judgment to decide where, and whether, the tool belongs.

The project’s researchers say they are already reaching out to schools, districts, organizations, and edtech companies interested in joining the next phase of work on AI governance, and they provide a contact email at research@edsurge.com.

generative AI education computational thinking AI literacy prompt engineering teacher use cases education governance edtech collaboration