84 Percent and Counting

Students aren't reckless, but schools aren't ready

Two-thirds of American high schoolers worry that using AI too much could make them “overly dependent” on the technology or “less intelligent.” This is not what the moral panic promised. The story was supposed to be about kids mindlessly cheating with ChatGPT, a generation of students outsourcing their thinking to machines. Instead, the data suggests something more interesting: teenagers seem to understand the stakes better than the adults wringing their hands about them.

The Numbers

AI use among students has surged. Between January and May of 2025, the percentage of high school students using generative AI for schoolwork climbed from 79% to 84%, according to College Board research. A RAND study published in 2025 found that 54% of students and 53% of English, math, and science teachers reported using AI for school—increases of more than 15 percentage points compared to one or two years prior.

The institutional response has not kept pace. UNESCO reports that only 10% of schools and universities currently have an official framework for AI use. The RAND study found that over 50% of teachers say their schools have no formal AI policy. Nearly one in five district leaders report having no policy at all, while more than a quarter leave decisions to individual teachers. Ninety percent of high school principals surveyed say they worry about teacher preparedness.

UNESCO has dedicated International Education Day 2025—this coming Friday, January 24—to artificial intelligence in education, calling for member states to invest in training teachers and students on responsible use. The timing reflects a growing recognition that adoption has outrun guidance by a significant margin.

What Students Actually Think

The College Board data complicates the familiar narrative. Two-thirds of high schoolers agree or strongly agree that using AI too much could make them overly dependent on the technology. A similar proportion worry it could make them less intelligent. As Fortune noted in December: “For all the handwringing about an ‘AI-fueled decline in learning,’ teenagers may actually understand the stakes better than adults think.”

This is not the profile of reckless adopters. Students are using AI widely and expressing caution about that use simultaneously. But in the absence of school-level rules, they are largely setting their own boundaries.

The institutional picture is uneven. A Hollywood Reporter piece from last week documented how elite Los Angeles private schools have shifted from panic to pragmatism. Buckley School now runs a $1,500 weeklong AI summer camp for middle schoolers. Harvard-Westlake, Windward, and Sierra Canyon have formed partnerships with AI education platforms and are training teachers on ChatGPT and Google Gemini. “Since day one the narrative in education has been really focused on cheating,” one educational consultant told the publication. “We’ve got to kind of go beyond that, I think, and meet the students where they’re at with the technology.”

Elsewhere, the picture is different—blanket bans, confused policies, teachers navigating alone.

The Longer Shadow

These same students are entering a job market being reshaped in real time. The anxiety about AI in the classroom connects to a larger anxiety about AI and work.

The World Economic Forum’s 2025 Jobs Report found that 41% of employers intend to reduce their workforce by 2030 due to AI. Entry-level and white-collar jobs—the traditional on-ramps for young workers—are among the most exposed categories. According to National University’s review of AI employment statistics, workers aged 18 to 24 are 129% more likely than those over 65 to worry that AI will make their job obsolete. Forty-nine percent of Gen Z job seekers believe AI has already reduced the value of their college education.

The WEF has introduced the term “AI precariat” to describe what may be coming: a class defined not just by economic insecurity, but by the loss of what work has traditionally provided. “When we are young, we dream about what we want to be when we grow up,” the organization wrote in August. “But in the age of AI, many of those dreams are being rewritten—or erased.” The concern is not only about jobs disappearing, but about what happens to identity and belonging when the anchor of work shifts beneath an entire generation.

Whether these fears prove warranted is an open question. A Yale Budget Lab analysis published in 2025 found that the occupational mix in the U.S. labor market is changing somewhat faster than during the early internet era, but not dramatically so—at least not yet. The displacement may be slower than the anxiety suggests, or the anxiety may be ahead of the curve.

The Picture

The pattern is consistent across the data. Young people are adopting AI rapidly, thinking about its risks seriously, and facing an uncertain labor market where the usual paths may not hold. The institutions meant to prepare them—schools, employers, policymakers—are moving slower than the technology. Some students are getting structured guidance and AI literacy training. Many are not.

What happens next is unclear. But the kids, it turns out, are already paying attention.