The Calculator Debate Is Back: What Schools Need to Get Right About AI

When handheld calculators started showing up in classrooms decades ago, a lot of educators were worried. The concern was understandable. If students could push a few buttons and get the answer, why would they bother learning how to do the math themselves? There was the assumption that calculators would make students lazy and replace learning with a shortcut.

We know how that story turned out. Calculators didn’t eliminate the need to understand mathematics. They became another tool. Students still learn arithmetic, algebra and calculus, but we no longer insist that every calculation be done with pencil and paper simply because that’s how previous generations learned.

Now we are having a remarkably similar argument about artificial intelligence. This time, the stakes are arguably higher. Generative AI can do much more than arithmetic. It can write an essay, summarize a book, explain a scientific concept, generate computer code, and work through a complicated problem in seconds. The concern isn’t simply that students will take a shortcut. It’s that they may hand over some of the thinking itself.

That concern deserves to be taken seriously.

The Cognitive Shortcut is Real

There’s a legitimate question about what happens when people rely too heavily on AI. Learning requires effort. We remember things because we wrestle with them. We develop writing skills by writing badly and then, with practice, getting better. We learn how to solve problems by getting stuck, making mistakes and figuring out why. Give a student an AI system that can immediately produce a polished answer and it becomes tempting to skip all of that uncomfortable work.

Some early research gives educators reason to be cautious. A 2025 randomized study involving university students found lower long-term knowledge retention among students who used ChatGPT as a study aid compared with students who used traditional study methods. A broader 2025 review of empirical research on generative AI in education similarly concluded that schools should ensure students have foundational knowledge before introducing AI, be taught AI literacy, and use the technology as a supplement to traditional learning rather than a replacement for it.

When teachers and parents worry that AI could become a cognitive crutch, I understand where they are coming from. But acknowledging that risk is very different from concluding that students should simply be kept away from the technology.

We’ve Seen This Movie Before

The calculator debate offers a useful lesson. The mistake wasn’t worrying about calculators. Educators were right to ask what skill students might lose by relying on calculators. In many jurisdictions, the mistake was keeping calculators out of the classroom.

Over time, educators gradually learned where the technology properly belonged. You teach the fundamentals first. Then you use the tool to do things that would otherwise be cumbersome or impractical. The calculator doesn’t replace mathematical understanding. Properly used, it extends what a student can do with that understanding. The calculator augments the student’s capabilities.

AI requires the same kind of distinction. A student learning to write an argumentative essay should know how to construct an argument before outsourcing one to an AI system. A student studying history needs enough knowledge to recognize when an AI-generated answer is wrong or oversimplified. A student learning computer science should understand what the code is doing rather than blindly copy-and-pasting whatever a chatbot generates.

But once those foundational skills are there, the educational question changes. How can AI help students explore an idea more deeply? How can they interrogate its answers? How can they find its mistakes? How can teachers use it to demonstrate bias, hallucinations and the difference between a plausible answer and a correct one? Or, more generally, how can AI be used to augment the student’s abilities?

Those are learning opportunities, too.

New York City is Testing the Boundary

New York City Public Schools is taking a particularly cautious approach during the 2026–27 school year. Student-facing generative AI is under a one-year moratorium from kindergarten through eighth grade. High school students have more limited, supervised access through approved programs and career-readiness coursework. At the same time, every high school student is required to complete AI-literacy instruction covering topics including how AI works, privacy, bias, and appropriate use.

The system is clearly concerned about students becoming overly dependent on generative AI, particularly before they develop foundational skills. Families and educators participating in the city’s public-feedback process expressed concerns that AI could replace some of the thinking that schools are supposed to teach. Students themselves expressed similar concerns about overreliance. These are all reasonable concerns.

At the same time, requiring AI literacy for high school students recognizes another reality: These tools exist, and students need to understand them. That’s where the most productive education debate should be.

AI Literacy is Literacy

Schools have always had to prepare students for the world they will enter, not the world their teachers grew up in (an all-too-often mistake). Today that world includes artificial intelligence.

Knowing how to use AI doesn’t simply mean knowing how to type a prompt into ChatGPT. Students need to understand that AI systems can produce convincing nonsense. They need to understand how bias occurs and how to recognize it. They need to know what happens to information they put into these systems. They must understand privacy and why it’s important. They need to distinguish between using AI as an assistant (augmenting their skills) and allowing it to do work they should do themselves.

Most importantly, they need to learn when not to use it. That requires education that touches many dimensions, including important life lessons such as responsibility, ethics, and professionalism.

Keeping students away from AI may temporarily prevent misuse inside a classroom. It doesn’t teach them how to use the technology responsibly when they encounter it everywhere else.

Teach the Student, Then Teach the Tool

There should be boundaries around the use of AI in education. A 9-year-old learning to write a paragraph is not in the same position as a 17-year-old learning how AI tools are being used in engineering or computer science. Different ages, subjects, and learning objectives call for different rules.

There will also be assignments where AI should not be used at all. If the purpose of the exercise is to determine whether a student can reason through a problem independently, having an AI system do the reasoning defeats the purpose. But that’s an argument for thoughtful instructional design, not for treating the technology itself as something students should never learn.

The challenge for educators is to create a symbiotic relationship between the student and the technology: preserve the difficult cognitive work that produces learning while teaching students how to use AI to augment their capabilities. That will require experimentation. Some approaches will fail. Teachers will need training. Schools will have to revise assignments and rethink how they assess what students actually know. All of those initiatives are hard, but they will bear fruit. That’s what it means to offer a modern, relevant, evolving education. Simply banning a tool saves a lot of effort but at what cost to the quality of education being offered?

Fifty years from now, I suspect the idea that students once debated whether AI belonged in education may sound a lot like the arguments we once had about calculators.

The important question is not whether students will use AI. It is whether we will teach them how to use it without surrendering the thinking they still need to do for themselves. In education, let’s think of AI not as artificial intelligence but as augmented intelligence: learning to use modern tools to augment students’ capabilities.