Just as school boards and administrators evaluate carefully every item placed in classrooms from the size of windows to furniture to whiteboards to textbooks, so too should the ubiquitous technologies used daily–nay, hourly–such as cell phones and laptops (including software) be rigorouly assessed. New York Times reporter Jessica Grose makes just that point in this article.

Educational technology in schools is sometimes described as a wicked problem — a term coined by a design and planning professor, Horst Rittel, in the 1960s, meaning a problem for which even defining the scope of the dilemma is a struggle, because it has so many interconnected parts that never stop moving.

When you have a wicked problem, solutions have to be holistic, flexible and developmentally appropriate. Which is to say that appropriate tech use for elementary schoolers in rural Oklahoma isn’t going to be the same as appropriate tech use in a Chicago high school.

I spent the past few weeks speaking with parents, teachers, public school administrators and academics who study educational technology. And while there are certainly benefits to using tech as a classroom tool, I’m convinced that when it comes to the proliferation of tech in K-12 education, we need “a hard reset,” as Julia Freeland Fisher of the Christensen Institute put it, concurring with Jonathan Haidt in his call for rolling back the “phone-based childhood.” When we recently spoke, Fisher stressed that when we weigh the benefits of ed tech, we’re often not asking, “What’s happening when it comes to connectedness and well-being?”

Well said. We need a complete rethink of the ways that we’re evaluating and using tech in classrooms; the overall change that I want to see is that tech use in schools — devices and apps — should be driven by educators, not tech companies.

In recent years, tech companies have provided their products to schools either free or cheap, and then schools have tried to figure out how to use those products. Wherever that dynamic exists, it should be reversed: Districts and individual schools should first figure out what tech would be most useful to their students, and their bar for “useful” should be set by available data and teacher experience. Only then should they acquire laptops, tablets and educational software.

As Mesut Duran — a professor of educational technology at the University of Michigan, Dearborn, and the author of “Learning Technologies: Research, Trends and Issues in the U.S. Education System” — told me, a lot of the technology that’s used in classrooms wasn’t developed with students in mind. “Most of the technologies are initially created for commercial purposes,” he said, “and then we decide how to use them in schools.”

In many cases, there’s little or no evidence that the products actually work, and “work” can have various meanings here: It’s not conclusive that tech, as opposed to hard-copy materials, improves educational outcomes. And sometimes devices or programs simply don’t function the way they’re supposed to. For example, artificial intelligence in education is all the rage, but then we get headlines like this one, in February, from The Wall Street Journal: “We Tested an A.I. Tutor for Kids. It Struggled With Basic Math.”

Alex Molnar, one of the directors of the National Educational Policy Center at the University of Colorado, Boulder, said that every school should be asking if the tech it’s using is both necessary and good. “The tech industry’s ethos is: If it’s doable, it is necessary. But for educators, that has to be an actual question: Is this necessary?” Even after you’ve cleared the bar of necessary, he said, educators should be asking, “Is doing it this way good, or could we do it another way that would be better? Better in the ethical sense and the pedagogical sense.”

With that necessary and good standard in mind, here are some specific recommendations that I’ve taken away from several discussions and a lot of reading. It’s unrealistic — and considering that we’re in a tech-saturated world, not ideal — to get rid of every last bit of educational technology. But we’re currently failing too many children by letting it run rampant.

At the State and Federal Levels: Privacy Protections and Better Evaluation

A complaint I heard from many public school parents who responded to my March 27 questionnaire and wanted a lower-tech environment for their kids is that they’re concerned about their children’s privacy. They couldn’t opt out of things like Google Classroom, they said, because in many cases, all of their children’s homework assignments were posted there. Molnar has a radical but elegant solution for this problem: “All data gathered must be destroyed after its intended purpose has been accomplished.” So if the intended purpose of a platform or application is grading, for example, the data would be destroyed at the end of the school year; it couldn’t be sold to a third party or used to further enhance the product or as a training ground for artificial intelligence.

Another recommendation — from a recent paper by the University of Edinburgh’s Ben Williamson, Molnar and the University of Colorado, Boulder’s Faith Boninger outlining the risks of A.I. in the classroom — is for the creation of an “independent government entity charged with ensuring the quality of digital educational products used in schools” that would evaluate tech before it is put into schools and “periodically thereafter.” Because the technology

is always evolving, our oversight of it needs to be, as well.

At the District Level: Centralize the Tech-Vetting Process

Stephanie Sheron is the chief of strategic initiatives for the Montgomery County Public Schools, the largest district in Maryland, and all the district’s technology departments report to her. She likened the tech landscape, coming out of the Covid-19 pandemic remote school period, to the “Wild West.” School districts were flooded with different kinds of ed tech in an emergency situation in which teachers were desperately trying to engage their students, and a lot of relief money was pouring in from the federal government. When the dust settled, she said, the question was, “Now what do we do? How do we control this? How do we make sure that we’re in alignment with FERPA and COPPA and all of those other student data privacy components?”

To address this, Sheron said, her district has secured grant funding to hire a director of information security, who will function as the hub for all the educational technology vending and evaluate new tech. Part of the standardization that the district has been undergoing is a requirement that to be considered, curriculum vendors must offer both digital and hard-copy resources. She said her district tried to look at tech as a tool, adding: “A pencil is a tool for learning, but it’s not the only modality. Same thing with technology. We look at it as a tool, not as the main driver of the educational experience.

At the Classroom Level: Ruthlessly Evaluate Every Tool

In my conversations with teachers, I’ve been struck by their descriptions of the cascade of tech use — that more tech is often offered as a solution to problems created by tech. For example, paid software like GoGuardian, which allows teachers to monitor every child’s screen, has been introduced to solve the problem of students goofing off on their laptops. But there’s a simple, free, low-tech solution to this problem that Doug Showley, a high school English teacher in Indiana I spoke to, employs: He makes all his students face their computer screens in his direction.

Every teacher who is concerned about tech use in his or her classroom should do a tech audit. There are several frameworks; I like the worksheet created by Beth Pandolpho and Katie Cubano, the authors of “Choose Your Own Master Class: Urgent Ideas to Invigorate Your Professional Learning.” In the chapter “Balancing Technology Use in the Classroom,” they suggest that teachers list every tech tool they are using and evaluate its specific functions, asking, “Are these novel or duplicative?” They also encourage teachers to write out a defense of the tool and the frequency of use.

I like these questions because they make clear that the solutions are not going to be one size fits all.

Students Deserve Authentic Connection

As I close out this series, I want to return to what Fisher said about the importance of student connection and well-being. Of course academic outcomes matter. I want our kids to learn as much about as many different topics as they can. I care about falling test scores and think they’re an important piece of data.

But test scores are only one kind of information. A key lesson we should have learned from 2020 and ’21 is that school is about so much more than just academics. It’s about socialization, critical thinking, community and learning how to coexist with people who are different from you. I don’t know that all of these are things that can be tracked in a scientific way, which brings me back to the idea of tech in schools as a wicked problem: These aren’t easily measurable outcomes.

Jeff Frank, a professor of education at St. Lawrence University, expresses a sense that I’ve had very well in a paper, “Sounding the Call to Teach in a Social Media Age: Renewing the Importance of Philosophy in Teacher Education.” He says students are “hungry for experiences that make them feel alive and authentically connected to other people and to deeper sources of value. Though filtering and managing life through technologies offers safety, predictability and a sense of control, it also leads to life that can feel extremely small, constraining and lonely. Teaching can offer a powerful way to pierce this bubble.”

Ultimately, I believe the only way kids will be able to find that deeper meaning is through human relationships with their peers and teachers, no matter how shiny an A.I. tutor appears to be at first blush.