After the past few weeks, there can’t be many people left still unaware of the rapidly growing power of artificial intelligence. No one knows quite what to do about it or what to expect and when. Parents fear for their children and how they might survive in this new world. But they should rather look to their children’s schools, where right under their noses, AI and educational technology – ed tech – is already damaging students’ futures and stealing their ability to focus, to learn and to think.
You only have to look the latest PISA results, released earlier this month, to see that academic performance in America is declining fast. The reading scores of our 15-year-olds have fallen by 14 points since the last test just three years ago – and if students don’t or won’t read, then there’s not much else they can learn. How and why has it come to this? Covid was once an easy answer, but the decline carried on as schools opened up again after lockdown. What did coincide with the slump in performance, though, was the arrival of new technology in schools, designed to engage already screen-addled children. What has accelerated the decline is widespread use of AI. The PISA results showed that children who do not use AI for schoolwork score, on average, about 20 points higher than students who do – the equivalent of being an entire school year ahead.
Laptops arrived in schools as a status good, not a pedagogical one
The original sin was the introduction of the Chromebook. I once watched a ninth grade student spend 45 minutes writing a single paragraph. The task wasn’t difficult. She was a capable writer. I knew she cared about the class and the assignment, but she was drafting on her school-issued Chromebook, the word processor only one of dozens of open tabs, each tempting – and ultimately succeeding – to steal her attention away from a half-written sentence. She’d type a few words, switch tabs, check the shared document her peers had converted into a makeshift group chat, respond with a joke, return to her essay, reread what she’d written, and start again. When the period ended, she looked exhausted, not from intellectual effort but from the constant labor of resisting distraction.
This wasn’t a failure of her discipline. It was a failure of the environment. She was tasked with writing an essay, but wasn’t given the conditions needed to write it. How did we end up here?
One-to-one device programs began in our independent schools. As early as the 1990s, and accelerating through the aughts, schools with an imperative to market their prestige cast laptops as a distinguishing feature. Their prospectuses advertised access to personal devices, interactive smartboards were shown off during open tours and “digital strategies” were placed front and center of principals’ welcome letters. Laptops arrived as a status good, not a pedagogical one.
At the time, what amounted to little more than a branding decision was accompanied by unsubstantiated claims about the importance of digital technologies in the classroom: “We’re preparing students for the future.” Which future? In what way were they being prepared for it? And according to what evidence?
Without much notice, the measure of a good school shifted from what came out of it to what was inside it. Clever marketing successfully convinced parents to select schools by assessing their resources rather than their outcomes. A cart of Chromebooks became legible proof of an institution’s professionalism, rather than a well-taught history class. It’s easy to show off pictures of students huddled around a laptop on your website. How do you do the same for sound ethical reasoning?
By the early 2010s, public schools followed suit. Some 23 million devices – laptops, tablets and Chromebooks – were purchased by American public schools in 2013 and 2014 alone. An arms race to equip students with the latest hardware had begun.
Eyeing the enormous potential of an untapped market, education tech software came next. From learning management systems such as Google Classroom to gamified math and literacy products such as i-Ready, apps quickly colonized the classroom. Today, the ed-tech industry is valued at more than $400 billion.
But what has all that money bought us? Not results, that’s for sure. Since one-to-one device programs became common, average PISA scores in all three categories – reading, math and science – have declined across participating countries. In 2015 the OECD reported that “there is little solid evidence that computer use among students leads to better scores in mathematics and reading.” Eight years later, UNESCO reaffirmed that conclusion, adding that mere proximity to a device was found to damage learning in 14 countries.
And yet, school prospectuses haven’t changed. Ed tech is everywhere. The Consortium for School Networking describes “nearly ubiquitous 1:1 implementation at every grade level” throughout the United States. Schools took up the mantra “move fast and break things” at exactly the moment they were the last institution on Earth that might exist beyond Silicon Valley’s reach.
In the classroom, this state of affairs has been a nightmare for teachers. From gaming to gambling – and yes, I mean gambling; sports betting apps are a fixture of teenage boys’ phones – students have a million and one opportunities to evade learning. Teachers have become reluctant officers, policing screens their own schools handed out, and suspicion is toxic in a profession that runs on trust. You’d be hard-pressed to find a high school teacher today who hasn’t caught a student watching Netflix, taking selfies, or shopping online during class. For many students, this is what the experience of being in school has become. Schools spend up to $500 on a Chromebook so Johnny can spend the better part of an hour setting his fantasy football roster during science class.
But the consequences extend far beyond wasted time. Once the machines were in the room, teachers found themselves competing against the best-funded behavioral engineering systems in human history. Rather than remove the competitors, the education profession decided to try to beat them at their own game.
Ed tech companies released applications that imitated the very techniques mobile games use to maximize their users’ engagement. Kahoot!, for example, is an incredibly popular interactive platform offering quiz games. Students scan a QR code to join, forcing them onto a device, and race to select the correct answer to multiple-choice questions that appear on screen. Along with colorful backgrounds and stimulating jingles, the platform prompts participants to give themselves nicknames and design their own cartoon avatars. At the end of each question, a leaderboard is updated, and the final results are announced by revealing a podium featuring the top three students’ animated characters. If that sounds a long way from a room of 15-year-olds arguing about whether appeasement or the Treaty of Versailles did more to cause World War Two, that’s because it is.
Once kids are hooked, ed tech companies benefit from harvesting their data
But students love Kahoot!. When its hair-raising jingle fills the room, kids who haven’t voluntarily raised their hands in weeks transform into the classroom’s most eager participants. “But they’re engaging. That’s great,” you say. Not quite. For some, anonymity is an opportunity to select lewd nicknames for a cheap laugh. For others, picking up their phones is a chance to catch up with their friends’ Snapchats. And for many, clicking random answers gets them ten minutes closer to the bell. What appears as engagement is in fact just a stew of trivial stimulation and wasted time. For teachers, it just necessitates more policing.
Kahoot! exists because our education system is suffering a crisis of engagement. How do you arouse your students’ interest when their minds have grown accustomed to high levels of digitally delivered dopamine? Scholar Maryanne Wolf calls it “cognitive impatience.” A brain habituated to skimming – consuming reams of short-form video, near-constant notifications and other rapid-fire digital stimuli – simply cannot do the slow analytic work that deep learning requires. Research backs up the point: just last year, a team from the University of California, San Francisco tracked the social media use of 6,000 children aged nine to 13. They found that even moderate use of social media, defined as just one hour per day, was associated with lower reading and memory scores.
Another team of researchers, north of the border at Toronto’s Hospital for Sick Children, conducted a 15-year study that found each additional hour of daily screen time in early childhood correlated with 9 to 10 percent lower odds of achieving a higher academic level in grade three reading and math and grade six math.
In 1953, American psychologist B.F. Skinner demonstrated that the variable-ratio reward schedule – in which a reward arrives, but unreliably – is the most effective method for producing predictable and persistent behavioral responses. Casinos call it “continuous productivity,” or a state of extended, near-uninterruptible play – the consequence of variable-ratio reward schedules applied to people. Pull-to-refresh, the infinite scroll and the simple notification that may or may not signal the arrival of something worth your attention all work on the same principle: there’s a reward you can’t predict.
Ed tech products such as Prodigy, a fantasy game in which children fight monsters by answering math questions, are rife with the same engagement mechanics. Kids who play the game at school are prompted to return after 4 p.m. to spin a prize wheel, encouraging them to move off of the free classroom version and on to a paid one at home. Once kids are hooked, ed tech companies benefit from harvesting their data, the most valuable commodity in the world. The Utah State Board of Education tested network traffic on 100 of its ed tech applications. They found that 52 percent of the apps collected data that their privacy agreements explicitly forbade, 61 percent shared student data with third parties, and 36 percent transmitted that data to advertisers.
But their incentives extend beyond data mining and lucrative school-board contracts. In exchange for convenience, programs such as Google Classroom, which organize homework and set reminders, ingrain a form of cognitive dependence. As the tech critic Cory Doctorow observes, a search company with 90 percent market share struggles to grow by finding new users, so it raises them to maturity instead: that’s quite literally Google Classroom.
The greatest casualty of the ed tech paradigm is the displaced capacity for sustained attention
Artificial intelligence extends the logic of cognitive dependence to all avenues of thought. I’ve seen students take a picture of a worksheet from their math class, upload it to ChatGPT, ask the AI to complete all the problems and copy the answers down in their own handwriting. Not a moment is spent thinking mathematically – or thinking at all, for that matter. The worksheet is completed, but nothing is learned.
Perhaps the greatest casualty of the ed tech paradigm is the displaced capacity for sustained attention. When one is used to superficial encounters with concepts and ideas, the ability to commit to anything past the point of immediate reward atrophies, and the tragedy of this is that almost everything worth learning is unrewarding before it is rewarding. If a student can’t endure the uncomfortable, he never reaches the reward at the other side. The range of what a school can plausibly offer him collapses because entire categories of experience sit beyond his capacity for engagement. Whole books become excerpts. Nuance cedes to imperatives. And education becomes instrumental. We haven’t narrowed what we teach; we’ve narrowed what can be received.
None of this requires stripping every device from every classroom. There are good reasons to teach with technology – namely those things that can’t be taught without it. Learning to code? Use a computer. Copying notes from the board? Use a pen. But the burden should sit with the screen – we must not ask why shouldn’t we, but why should we. That’s the default we abandoned some years ago without giving it much thought.
The current trajectory of educational technology is not inevitable. The declining scores, the erosion of attention spans, the increasing anxiety and decreased resilience – these are consequences of specific choices that can be changed. Schools can be sanctuaries. They can be places where children are protected from pernicious and extractive business models. They can be the level ground that provides youngsters with the conditions we know they need to thrive: embodied engagement with others, productive struggle and the space to think clearly.
Simone Weil wrote that “attention is the rarest and purest form of generosity.” We have spent at least 15 years degrading our students’ capacity for it.
Comments