AI use reaches 83% of U.S. teens as school guidance struggles to keep pace

Britebound’s national survey of 3,000 students finds widespread AI use alongside major gaps in teacher guidance, school support and perceived workforce readiness

Britebound surveyed 3,000 U.S. students in grades 7-12 about their use of AI, school guidance and preparation for future work

AI is already routine for most U.S. teenagers, but far fewer believe their schools are preparing them to use it well.

A national survey from Britebound found that 83% of students in grades 7-12 use AI tools, with around two-thirds of users turning to them at least a few times a week and 30% using them every day. Yet just 36% of teens say their school is teaching them enough about AI, while only 42% believe they have the AI skills they will need for a future job.

That gap between adoption and preparation sits at the center of Britebound’s Teens and the AI Future research, based on an online survey of 3,000 U.S. teenagers aged 13-18 enrolled in secondary school. Demographic quotas and weighting were used to align the sample with national benchmarks.

The findings also come with an important limitation. Britebound describes the relationships in its data as associational, not causal. The survey can show, for example, that students who report greater teacher encouragement also report different patterns of AI use, but it cannot establish that one produced the other.

What it does show is how unevenly the current school experience is developing. Only 21% of teens say most of their teachers encourage AI use for schoolwork. Another 37% describe a mixed environment, where some teachers encourage it and others do not, while 19% report discouragement and 17% say AI use is strictly prohibited.

Bans are not keeping AI out of schoolwork altogether. Among teens who say their teachers strictly disallow AI, 43% still report using it for school assignments.

AI use changes when teachers are involved

How students use AI also differs according to the school environment they describe. Among students using AI for school, seeking answers to homework problems is the most common activity, reported by 61%. But substantial numbers also use it to generate ideas, at 52%, explain difficult concepts, at 51%, edit writing, at 46%, summarize texts, at 45%, and create flashcards or practice questions, at 39%.

Students whose teachers mostly encourage AI report a broader pattern of use than those in restrictive environments. Answer-seeking is 12 percentage points lower among students in encouraging environments, while every other school-related use measured by Britebound is higher. The largest differences are in creating flashcards or practice questions, 25 points higher, and drafting papers, 19 points higher.

That does not prove teacher encouragement causes students to use AI differently. Schools with different approaches to AI may also differ in resources, training, technology access and other characteristics. But simply banning the technology does not appear to remove the behavior schools may be most concerned about. Students are still using AI, and in more restrictive environments that use is more heavily concentrated around getting answers.

The divide becomes even clearer when Britebound looks at how prepared teenagers believe they are. Researchers grouped respondents into four readiness profiles based on their reported understanding of AI, confidence in having the AI skills required for future work, and whether they believe their school is teaching them enough.

Only 24% fell into the “High AI-Ready” group. Another 39% were classed as “Mid AI-Ready”, while 19% were “Low AI-Ready” and 18% were unsure. Among teens who say most teachers encourage AI use, 61% fall into the High AI-Ready group. For students whose teachers strictly prohibit AI, that falls to 9%.

Private and public school students report very different AI experiences

Some of the widest gaps run along school type. More than half of private school teens, 55%, say they use AI every day, compared with 21% of students in traditional public schools. Just 5% of private school students rarely or never use AI, against 19% in traditional public schools.

The gap in teacher encouragement is larger still. Some 57% of private school students say most of their teachers encourage AI for schoolwork, compared with 13% of students attending traditional public schools. And while 56% of private school respondents are classed as High AI-Ready, that drops to 19% among traditional public school students.

Parental education follows a similar pattern. Teenagers whose parents do not hold a four-year degree are twice as likely to report little or no AI use as those whose parents have at least a four-year degree, 22% compared with 10-11%. High AI readiness is also substantially more common among teenagers whose parents hold advanced degrees.

Britebound identifies disparities by race and ethnicity in its own survey too, although it is careful to note that findings on AI use vary between national studies. In its data, White and Asian teens are more than twice as likely to fall into the High AI-Ready group as Black and Hispanic/Latino teens.

The organization argues that the emerging divide is therefore less about whether students can access AI at all, and increasingly about whether they are getting the guidance and opportunities to use it for learning and skill development.

Teens expect disruption, but many still do not know what AI means for their future

Teenagers are also far from settled on what AI will do to jobs and society. Two-thirds, 66%, expect AI to create new kinds of jobs. At the same time, 60% believe many people will lose existing jobs and 57% think AI will make it harder to get a job.

There is a similar tension elsewhere in the responses. While 51% believe AI will make jobs easier, the same proportion expect people to have more leisure time, and 51% say they are concerned about AI.

When Britebound grouped responses to nine questions about the future, uncertainty was the largest category by some distance. Some 45% of teens fell into its “Uncertain” group, compared with 19% classed as “High-Risk”, 16% as “High-Upside” and 19% as “Ambivalent”, meaning they held strong expectations of both benefits and risks.

More exposure did not simply make teenagers more optimistic. Among students whose teachers mostly encourage AI, 53% were in the Ambivalent group, suggesting they were more likely to identify both opportunities and disruption. Only 20% were classed as Uncertain. For students in schools where teachers prohibit AI, the picture flips: 47% were Uncertain, 31% High-Risk and just 7% Ambivalent.

Britebound places those findings against a school system where teacher training and formal AI policy are still developing. Research cited in the paper found that 68% of K-12 public school teachers surveyed by Gallup and the Walton Family Foundation in 2025 had received no school or district-provided AI training, while only 19% said their school had an AI policy.