There was a third-grade class in Massachusetts where a child got a “Certificate of Completion” for playing a drag-and-drop computer game that had almost nothing to do with AI. The certificate, which was made in collaboration with Amazon, was written in language that people would understand. Most honest people would say it was a branding exercise. It was in a backpack on its way home before any parent had a chance to say anything.
Something has been quietly building up in American public schools for a few years now. That moment, which was small and almost funny, shows it. Instead of big policy announcements, AI tools are coming into classrooms in the form of Chromebooks, reading bots, slide show prompts, and certificates of completion. The software is already on the computer when the parents find out about it.
In March, the New York City Department of Education put out its first guidelines for how AI can be used in public schools. It has a traffic light system: red means AI can’t be used to grade or decide a student’s academic path; yellow means AI needs to be used with care; and green, it seems, means go ahead. More than 900,000 students are covered by the framework. It led to a seven-hour public meeting where parents were angry in a way that didn’t seem like panic but rather like exhaustion. Like when you realize you’ve been asking the wrong questions for too long and are now tired of it.
A number of parent and community groups are now calling for a two-year ban on AI use in public schools in New York. Liat Olenick, co-founder of Climate Families NYC, said that the city’s rules were not good enough because they didn’t do enough to limit student use and didn’t take into account concerns about everything from environmental damage to kids‘ brain development. Before any more tools are used, her group wants neuroscientists, climate scientists, and education researchers to be involved in making the rules. It makes sense to do that. Also, the fact that the tools are already running makes things more difficult to plan.
The teachers’ union NYSUT wants limits on screen time and AI contact that are appropriate for kids’ ages. One idea is to stop kids younger than second grade from interacting directly with AI. It was made clear by the president of the union: teachers are not against technology; they are for kids. It’s important to think about this difference for a moment because the tech industry has done a good job of mixing these two points of view, making it seem like opposing AI in education is the same thing as opposing progress.
In all of this, Google has a certain role to play. Because they were cheap and easy to set up, Chromebooks became the standard for remote learning in many school districts during the pandemic. By the end of 2020, Chromebook sales had jumped by a huge amount year over year. This created a captive market for Gemini, Google’s set of AI tools. If a sixth-grader opens a blank document today, they will see the message “Help me write.” Before she gives a speech, it says, “Help me visualize.” The tools don’t need to be asked to join. They keep going.
People who want to use AI in K–12 schools say that kids who learn about computers early on become more tech-savvy, that adaptive learning tools help teachers learn more about each student, and that AI will be a part of most jobs in the future. The points made are not bad. Some of them might even be right. But there is a big difference between getting kids ready for a world with AI and giving a third-grader an Amazon-branded certificate before any teacher has considered what that teaches.
The practical view came from Naveed Hasan, a computer programmer, parent, and member of New York’s Panel for Education Policy. He said that schools should be able to set the terms because they are sometimes the only customers that certain vendors have. The kids’ needs should come before the vendor’s way of making money. Also, that seems pretty clear. Still, as I see this happen in many cities and school districts, I can’t help but notice how rarely that principle actually guides the choices.
Late in May, New York’s Chancellor said that his office had “missed the mark” when telling families about the framework. It is planned to give more detailed advice later this year. We don’t know yet if it will answer the more important questions about cognitive development and what it means for a child to be able to solve a problem without a chatbot finishing the thought. Technology changes faster than the rules that are made. There is a lot of trouble in that gap.
