In 1956, an ancient gentleman named Samuel James Seymour appeared on a game show called “I’ve Got a Secret,” where he revealed, in a small voice weakened by a recent fall, that he was likely the last living witness to the assassination of Abraham Lincoln. The blown-out kinescope transmission of the episode, preserved on YouTube, is a wormhole—a shortcut through the space-time continuum. The viewer takes this shortcut via her laptop or phone, through a television program aired within living memory, featuring a guest who entered adulthood not yet knowing cars or planes or skyscrapers or light bulbs, and whose own father was born before germ theory and the sewing machine.
I often feel as though I am shuffling onto the stage of my children’s education in the manner of Mr. Seymour. On one side of the wormhole is the stuff of the public K-12 schools I attended in the nineteen-eighties and nineties: chalkboards and Scantrons, film strips and secondhand Signet Classics, and big, battered textbooks in grocery-bag slipcovers, passed down from student to student, year upon year. On the other side of the wormhole are my kids’ Chromebooks, pre-loaded with Google’s suite of education software and A.I. tools, along with apps from myriad other companies. I cup my hand to my ear and ask the kids to speak up as they recite the gamified learning platforms they have known: Wayground (né Quizizz), BrainPOP, ABCya!, Epic!, Kahoot!. My son, who is nine, does homework games on the free version of a cartoon show called 99math, but he’d like to join the app’s “Premium Club,” whose members can “Play UNLIMITED GAMES at home,” “Collect RARE EGGS and get EPIC PETS,” and “Get EPIC OUTFITS for your avatar,” all for just $9.99 a month.
My daughter, age eleven, spends a chunk of every social-studies class on Blooket, where she earns tokens by taking multiple-choice quizzes arranged by subject and grade level. (A recent Veterans Day–themed Blooket quiz led off with “Veterans give up a lot to serve their countries and show a lot of _____.” The correct answer is patriotism.) Using her tokens, she has assembled an army of cartoon avatars known as Blooks; her most prized platoon of late is her “Spooky Pack” of ten Blooks—a googly-eyed creature, a creepy caramel apple—who look like extras from a factory-outlet remake of “Monsters, Inc.” She acquired them ahead of Halloween by answering sub-grade-level questions about the water cycle. I bleat hoarsely at her through the wormhole, demanding to know how the avatars and the coins and all the clicking help her learn anything. “This is the not-learning part,” she explains, gently. I feel my dentures loosening as I wander away.
On my own computer, I click around in hapless search of a university-level tutorial or peer-reviewed study that might explain the cognitive benefits of Zorblitz or Pirate Pool or Fishing Frenzy (three of Blooket’s many game modes). Instead, I stumble on a promotional video for yet another platform familiar to my kids—a program called Boost Reading. In Boost, instead of clicking answers in order to earn Blooks, you click answers in order to earn accessories for avatars known as Curiosos. “The kids are really engaged,” a literacy coach murmurs despondently in the video, over a closeup of a student wearing headphones and staring at a screen. A fourth grader tells the camera, “I feel as if I’m not even reading. I feel as if I’m just playing a game.” She means it as a compliment. Some ersatz Philip Glass tinkles in the background, the sad-anxious sound of a mental-health P.S.A., or of your parents’ divorce, or of an entire nation sliding into an oddly Pavlovian illiteracy.
Boost Reading is one of the many products offered by the ed-tech company Amplify Education, which is backed by hundreds of millions of dollars in venture capital. These V.C. firms include Learn Capital, which, according to its website, focusses on “AI-powered learning, digital education, workforce training, and career advancement”; and A-Street Ventures, which centers “digital-first instructional materials in curriculum and new paradigms for student assessment.” In New York City Public Schools—the largest system in the country, with nearly nine hundred thousand students and an annual budget of about forty-five billion dollars—Amplify has been a major presence for more than fifteen years. Joel Klein, who was schools chancellor under Mayor Michael Bloomberg and a key national force behind the reorientation of public schools around ed-tech, later spent several years as chief executive of Amplify. In fiscal year 2026, Amplify received $12.4 million in contractual payments from the city’s Department of Education, including for curricula in math, science, and English Language Arts (E.L.A.).
The ed-tech that has swallowed our public schools is overwhelmingly driven by private capital, and Amplify is just one example. In fiscal year 2026, New York City paid contracts totalling $25.6 million to Curriculum Associates, whose majority stakeholders are private-equity firms. The company’s flagship product is i-Ready, a reading-and-math-assessment software platform that has kept its foothold in New York City and across the country despite being universally reviled by students, parents, and teachers. Curriculum Associates is facing multiple legal challenges related to i-Ready, including a class-action lawsuit, filed in the District of Massachusetts, alleging violations of consumer-protection and privacy laws; and a false-advertising case, filed in D.C. Superior Court, brought by an education nonprofit. The latter suit alleges that i-Ready “does not tell a teacher what question a child missed or why they missed it”; that the software promises “personalized instruction” but instead “runs students with opposite needs through the same pre-built lessons”; and that it measures students’ progress “on Curriculum Associates’ own test of its own lessons, scored by Curriculum Associates—a closed loop its marketing nowhere discloses.” (Curriculum Associates has denied any wrongdoing.)
Public schools would seem an odd fit for private equity, which is typically premised on a strategy of slashing costs, trimming head count, and finding the right moment to sell up. Yet Heggerty, which provides literacy curricula to some New York City schools, is backed by private equity. The majority owner of Really Great Reading (E.L.A. software) is private equity. The sole owner of Savvas (software for multiple subjects) is private equity. And one of the D.O.E.’s top curriculum venders, H.M.H.—a descendant of Houghton Mifflin, a name synonymous with old-school textbooks—is owned by the private-equity firm Veritas Capital. H.M.H. and its subsidiary N.W.E.A. have a total of nine approved learning programs in New York, including the E.L.A. curricula “Into Reading” and “Into Literature,” which are used in two-thirds of the city’s districts. In fiscal year 2026, the D.O.E. made payments exceeding twenty million dollars to H.M.H.
Veritas Capital, which has significant A.I.-related investments, refers to H.M.H. on its website as an “adaptive learning company,” though H.M.H.’s own site is more explicit on this point: “Only HMH puts AI intelligence inside the high-quality instructional materials—by design, not an add-on.” This is a surprising bona fide for the most prominent curriculum provider in New York City, where, in response to intense anti-A.I. public sentiment, Mayor Zohran Mamdani and his schools chancellor, Kamar Samuels, recently announced a one-year moratorium on A.I. for students in eighth grade and under, along with recommendations on reduced screen time.
Private capital’s entanglement with increasingly A.I.-driven ed-tech places it at cross-purposes with a growing number of public-school advocates. The D.O.E. has imminent plans to launch five A.I. pilot programs for high-school students, in math, writing, and A.I. literacy; two of the pilots are from nonprofits that have received significant funding from Google, a third is from a company backed by venture capital, and a fourth is from Intel, whose top shareholders include the A.I. behemoth Nvidia. But the AIM Coalition—the umbrella organization of teachers, families, and advocates that successfully agitated for the A.I. moratorium—is militating against the pilots, arguing that they have no proven educational benefits, that they may harm teen-agers’ cognitive development, and that they effectively compel teachers and students to supply A.I. companies with free labor and personal data.
The incursion of A.I into classrooms is so recent, so freshly strange, that it may yet be possible for advocates to unwind it further—and perhaps unwind a lot of unwanted and pernicious ed-tech along with it. But it is also terribly strange that so many avatars of venture capital and private equity have been permitted to financialize public schools, which are neither designed nor intended to create profits or customers. It’s stranger still that these investors see so many dollar signs in a system such as New York City’s, where, according to the D.O.E.’s estimations, more than seventy per cent of students live in poverty.
Public schools have been providing private companies with straightforward revenue streams for generations; the outfit now known as H.M.H., for example, has existed in one form or another since the early nineteenth century. In the olden days, though, a school in the market for curricula could simply offer money to a publisher in exchange for its textbooks, then use them until they fell apart. The publisher was not granted access to the personal information or schoolwork or grades of the children using the books, as is often the case with Google and Microsoft’s learning-management platforms. These purchases did not put the school on the hook for future software or tech upgrades. And, unlike apps, books do not disappear overnight if a school does not renew its subscription to them.
The technologization of learning, in short, introduced elements of technofeudalism into the relationship between schools and ed-tech companies, creating far more opportunities for monetizing students by harvesting their data. And no company plays the part of feudal lord better than Google, which earlier this year foisted its Gemini suite of A.I. tools on public schools, much to the surprise and dismay of many educators and families. Last week, Google announced that early next year, by popular demand, it would permit schools to turn off A.I. features in its K-12 software, but that schools choosing to do so would also lose access to Google Search—a trade-off that carried a whiff of extortion. Of course, students in search of information can go directly to the URLs of trusted sources or use an alternative search engine such as DuckDuckGo, as Naveed Hasan, a computer scientist who has two children in public schools in Manhattan, told me. “This is actually a great opportunity to say, ‘Hey, we don’t need most of this Google stuff,’ ” he said. “Somehow they’re trying to blackmail us with a free product. Like, who cares?”
Technofeudalism also perpetuates itself in the form of the computer-based assessment. Though standardized tests—preparing for them, administering them—have been eroding classroom-instruction time across the country since the signing of the No Child Left Behind Act, in 2002, their corrosive effects intensified in many districts after COVID. In 2021, as schools in New York City were reopening for full-time, in-person learning, the D.O.E. began mandating triannual online assessments in reading and math. These new tests, intended to measure pandemic-driven learning loss, were sourced from MAP Growth, which is developed by N.W.E.A.; i-Ready, from Curriculum Associates; and Acadience Learning, whose parent company is backed by private equity. “The logic was, ‘We need to identify the gaps if we’re going to close them,’ ” Kyle DeAngelis, who is a third-grade teacher at a public school in Harlem, told me. “That made sense after a global-health meltdown. But, six years later, we’re still doing the same things, and schools are not getting the resources they need to close those gaps.”
Hasan is one of five elected parent representatives to the Panel for Educational Policy, the governing body that helps determine rules and budgets for the D.O.E. “All of our curriculum now is based on test scores,” he told me. “The standardized tests don’t demand long-form reading or writing, and therefore the kids aren’t taught to have the endurance to read a book.” Only about fifteen per cent of public schools in New York City currently employ a librarian. “They have mandated a curriculum with basically no books in it,” Hasan said. “We don’t have libraries and librarians because there’s no test to see if your kid is good at library.”
So far, the devil’s bargain of exchanging books for test scores appears to be a one-way transaction. According to the most recent test data from the National Assessment of Educational Progress, the reading scores of both fourth and eighth graders in New York State have dropped nine points in the past decade. Those results are especially bleak when you consider that the tests don’t measure for a student’s reading stamina, or quantify the joy and excitement a young person might feel in reading a book for fun, as opposed to scanning an excerpt in order to extract the correct answer and maybe a prize of some kind: a few digital coins, a new hat for her avatar, a robot’s automated words of praise.
Given the expense and inefficacy of ed-tech, the D.O.E.’s outsourcing of curricula is especially puzzling to many teachers and parent advocates. “Why are we reaching out to private industry for what should be core public responsibilities?” Hasan asked. “Curriculum design should be done in-house, by teachers who know their local communities. Otherwise, the priorities will be set by the venders.” Hasan and other advocates singled out H.M.H.’s “Into Reading” program—which is widely disliked for its reliance on excerpts, lack of diverse authors, and flat, voiceless content—as illustrative of the need for local curriculum design.
DeAngelis, who has a master’s degree in curriculum and teaching, said he was concerned that H.M.H.-style heat-and-serve learning may have a de-skilling effect on his youngest colleagues. “With the proliferation of these prepackaged curricula, I fear we are raising a generation of teachers who believe that teaching is downloading a lesson and just delivering whatever is in there,” he told me. De-skilling is likewise a concern in any school system, including New York’s, where inexperienced teachers are permitted or even encouraged to use A.I.—arguably a form of cognitive outsourcing—to generate lesson plans, classroom presentations, and feedback on student work. Martina Meijer, who teaches fourth grade in Brooklyn, told me that teachers are frequently receiving A.I.-written feedback from their supervisors. “It’s the same quality of work that a middle-school teacher would get from a student essay written by A.I.,” she said.
Meijer and DeAngelis are both part of the Movement of Rank-and-File Educators (MORE), which is the social-justice caucus of the United Federation of Teachers and a member group in the AIM Coalition. MORE-U.F.T.’s adversarial approach to ed-tech and A.I. in schools diverges sharply from the incautious optimism of the mainline U.F.T., which represents most public school teachers in New York City. Last year, the U.F.T. and its counterpart union, the American Federation of Teachers, launched the National Academy for AI Instruction, a training program that is funded by a twenty-three-million-dollar partnership with Microsoft, OpenAI, and Anthropic. The academy’s classes and online resources offer guidance on how teachers can harness A.I. to lighten their workloads; in one course, teachers can “practice crafting AI prompts specifically for English learners,” and, in another, they can “learn how to build and personalize AI agents that act as instructional aides, planners, or feedback partners using educator-friendly AI tools.”
In accepting A.I. money for A.I. purposes, the teachers’ unions seem to be instantiating an idea that many of its rank-and-file members reject: that A.I. and ed-tech writ large can aid and enhance the work of teachers, as opposed to eroding or supplanting their role. “There’s no analysis of the labor issues that are raised by accepting corporate curricula and the rolling out of A.I. to address unacceptable labor conditions,” Meijer told me. “If the situation is that teachers are training computers to do their job for them, then we are at a very weak position at the bargaining table.”
The situation that Meijer describes is one that many D.O.E. leaders may have tacitly accepted. Richard Carranza, who was chancellor from 2018 until 2021, went straight from that job to an executive position at the ed-tech company I.X.L. Learning. Samuels’s immediate predecessor as chancellor, Melissa Aviles-Ramos, is now an “executive in residence” at H.M.H. And Miatheresa Pate, currently the D.O.E.’s chief academic officer, is the recipient of a fellowship jointly offered by Google and G.S.V. Ventures, an ed-tech venture-capital firm.
Yet the alternative vision for public education presented by members of AIM and MORE-U.F.T.—one that moves away from screens and centers books, pencil-and-paper learning, and a sense of being rooted in a physical place, a community—is less costly, more effective, and more human than the buffering in-app technofeudalist nightmare that our children currently occupy. Anyone who is above a certain age or able to read a D.O.E. spreadsheet can understand this, and can grasp how easy it would be to return to this model of pedagogy. Anyone can climb through the wormhole.
The sole major hurdle, it seems, is that private capital does not see this model as scalable or high-R.O.I. This shouldn’t be a hurdle at all, except that many public-education leaders in New York and elsewhere “have bought into the future outlined by tech and finance,” Craig Garrett, a parent leader in the AIM Coalition who lives in Brooklyn, told me. “The only thing that has worked so far is political pressure. If our classrooms ever return to textbooks and Scantrons, it will happen because the political price of standing up to thousands of parents, students, and teachers became too high.” ♦