The National Assessment of Educational Progress released its long-term trend results on Wednesday, June 10, comparing today's students against the test's original 1970s cohort. Nine-year-olds are closing the pandemic-era gap, posting reading and math scores close to pre-2020 levels. Thirteen-year-olds are not. Their average reading score has fallen seven points since 2012, and their current performance is statistically indistinguishable from thirteen-year-olds tested decades ago, before the internet, before smartphones, before any of the interventions schools have layered on since. Math scores for the same age group fell fifteen points over the same period.
NAEP's own data points to a contributing habit, not just a skill gap. In 1984, thirty-five percent of thirteen-year-olds said they read for fun almost every day. In the 2025 assessment, fourteen percent said the same, a number that has been sliding for over a decade and shows no sign of reversing. A score is a snapshot. A habit is the mechanism that produces the score, and the mechanism is the one number on this report that a teacher can act on directly without waiting for a curriculum office to respond.
The Senate HELP Subcommittee on Education and the American Family held a hearing this week titled "The Future of K-12 Education in the Age of Artificial Intelligence," its first session on the subject. Chairman Tommy Tuberville of Alabama said, "Artificial intelligence is changing the world our kids are growing up in, and whether you like it or not, AI is going to be part of their education, their careers, and their daily lives." Ranking member Lisa Blunt Rochester of Delaware cited figures from Microsoft and Pew Research showing eighty-four percent of students and eighty-five percent of teachers already use AI in some form. Tuberville also announced that he and Blunt Rochester requested a Government Accountability Office study on the effects of AI in K-12 education, and the GAO has agreed to open the review. Witnesses raised student privacy, output accuracy, effects on critical thinking, and access gaps for rural districts as the four issues any federal rule would need to address.
A hearing is not a rule, and nothing here changes what you can do in your classroom tomorrow. What it signals is that the adoption numbers Blunt Rochester cited, eighty-four percent of students and eighty-five percent of teachers, have outrun any federal framework for how that use should be governed. Your district's AI policy, whatever it currently says, was written without the GAO review Tuberville just requested. That review will take months. Your students are not waiting on it.
The same NAEP long-term trend release that showed thirteen-year-olds' reading scores stagnant also tracked how often students read outside of assigned work. In 2012, twenty-seven percent of thirteen-year-olds reported reading for fun almost every day. In the 2025 data released this month, that figure stands at fourteen percent, a decline researchers and reporting from Chalkbeat and NPR connect directly to the score plateau. Nine-year-olds show a smaller version of the same pattern, down from fifty-three percent in 2012 to thirty-seven percent now, but their scores are still recovering toward pre-pandemic levels. Thirteen-year-olds, who lost more instructional time during the pandemic at a more consequential developmental stage, are not.
NAEP does not assign a cause, but the correlation is consistent across both age groups: as voluntary reading drops, score growth slows or disappears. You cannot fix a national habit, but you can fix what happens in your room for forty-five minutes a day. A score gap that opened over a decade will not close with a single unit. It closes, if it closes at all, the same way it opened: incrementally, through what students do with text when no one is grading it.
Catalina Foothills School District in Tucson surveyed teachers, students, and parents after one full year under its "off and away" phone policy, and AZPM reported the results this week. At the high school level, ninety-seven percent of parents agreed with the policy, and two-thirds of students called the current expectations reasonable. Nearly eighty percent of high school teachers said they would support an even more restrictive policy and reported improved peer interaction over the year. The middle school numbers told a more divided story: more than seventy percent of teachers reported improved student-to-student interaction, but only about one in four students agreed they were interacting more with peers.
The gap between what teachers observed and what students reported is the finding worth sitting with, not the headline support numbers. Adults in the building are measuring hallway conversation and classroom focus. Students are measuring something else, possibly a loss they don't have language for yet, possibly nothing the survey was built to ask about. A policy can succeed by every metric adults track and still feel like a loss to the people living under it. Both things can be true in the same building.