AI Homework Use Jumped 14 Points in 7 Months
RAND's American Youth Panel shows student AI use for homework rose from 48% to 62% in seven months, while critical thinking concern rose too.
Anshika Verma · Founder, GrowWise · Last updated
RAND's American Youth Panel found that the share of students (ages 12-29) using AI for homework rose from 48% in May 2025 to 62% in December 2025 - a 14-point jump in seven months. Over the same period, the share believing AI harms critical thinking rose from 54% to 67%. Adoption and anxiety are climbing together, in the same population, at the same time.
AI Homework Use Jumped 14 Points in Seven Months. So Did the Fear. Something Has to Give.
Two numbers from the same RAND survey, seven months apart, tell a story that neither one tells alone.
Students using AI for homework: 48% in May 2025 → 62% in December 2025.
Students who believe AI harms critical thinking: 54% in May 2025 → 67% in December 2025.
This is RAND's American Youth Panel, surveying more than 1,000 respondents ages 12–29 enrolled in school during the 2025–26 year, published March 2026. Both numbers moved up, together, fast.
The tension that should worry every school leader
Usually, when people adopt a technology more, their concern about it settles - familiarity breeds comfort. That isn't what's happening here. Usage rose 14 points and concern rose 13 points in the same window. Whatever is driving adoption is not resolving the worry that comes with it.
There's a sharper detail buried in the data: 78% of students who don't use AI express concern about its effect on critical thinking, compared to 60% of students who do use it. The people most worried are disproportionately the ones abstaining. That's not the profile of a moral panic dying down as people get comfortable - it's the profile of a technology whose heaviest users have made a conscious trade they're not entirely sure about, while the people watching from outside see the trade more clearly.
Female students report higher concern (75%) than male students (59%), which is worth a school's attention on its own, independent of the AI question.
Why this can't stay stable
A population where both the behavior and the anxiety about that behavior are rising together is not in equilibrium. Something eventually snaps - either toward much wider, uncontrolled use, or toward the kind of policy correction NYC and LA Unified just made.
Given what happened in September 2026 at the two largest school districts in the country, the snap has already started, and it went toward restriction, not expansion.
What schools do with this, practically
If 62% of students are already using AI for homework by the time your school reads this, a policy that assumes homework is untouched, uncoached, and student-authored is already fiction for most of your students. That doesn't mean giving up on homework as a signal. It means schools need a second, harder-to-fake source of evidence about what a student actually understands - something that happens in the room, under the teacher's own observation, and gets diagnosed at the level of the specific skill rather than a single grade.
That's the gap teacher-facing instructional intelligence is built to close: not detecting whether a student used AI on last night's worksheet, but generating a precise, current picture of what each student actually knows from the assessment evidence a school already produces - evidence a chatbot can't touch, because the AI never interacts with the student in the first place.
See Teacher-Facing AI in Action.
Related reading
FAQ
How much did student AI use for homework increase in 2025?
From 48% in May 2025 to 62% in December 2025, a 14-percentage-point increase in seven months, per RAND's American Youth Panel.
Are students who use AI more or less worried about its effects than non-users?
Less worried. 60% of AI users express concern about critical thinking effects, compared to 78% of non-users - the people abstaining are more concerned than the people using it.
Is concern about AI's effect on critical thinking rising or falling?
Rising. It went from 54% to 67% over the same seven-month period that usage rose, moving in the same direction as adoption rather than the opposite.
What AI tools are students using most for homework?
Chatbots lead at 60% overall use; ChatGPT is the single most-used tool at 53%. Google Gemini use more than doubled over the survey period, reaching 28%.
Sources
Frequently asked questions
- How much did student AI use for homework increase in 2025?
- From 48% in May 2025 to 62% in December 2025, a 14-percentage-point increase in seven months, per RAND's American Youth Panel.
- Are students who use AI more or less worried about its effects than non-users?
- Less worried. 60% of AI users express concern about critical thinking effects, compared to 78% of non-users - the people abstaining are more concerned than the people using it.
- Is concern about AI's effect on critical thinking rising or falling?
- Rising. It went from 54% to 67% over the same seven-month period that usage rose, moving in the same direction as adoption rather than the opposite.
- What AI tools are students using most for homework?
- Chatbots lead at 60% overall use; ChatGPT is the single most-used tool at 53%. Google Gemini use more than doubled over the survey period, reaching 28%.
See teacher-facing AI in action
AI that helps teachers see misconceptions and decide what to reteach next, without student chatbots.
2026 AI-in-schools policy comparison · Five questions before signing an AI vendor