Jobs for the Future surveyed 3,020 people aged 16 and older about AI use in education and work. The findings reveal a significant guidance gap: while AI use is surging, students are largely self-directing their learning.

• Use is up: 70% of learners use AI daily or weekly for education, up from 59% in 2024
• Integration growing: 69% say AI tools are incorporated into lessons, up from 57% in 2024
• But guidance lags: 48% use social media for AI help vs. only 23% who ask their schools
• Self-directed: 46% learn AI by experimenting on their own, 44% use YouTube or informal courses

“Learners are not sitting on the sidelines of the AI transition. They are actively experimenting, adapting, and integrating these tools into their educational experiences. But increased use does not automatically translate into increased efficacy.” — Ben Pring, JFF Center for Artificial Intelligence & the Future of Work

The Honest Take

This is both encouraging and concerning. Students are proactively learning—but they’re learning from TikTok, YouTube, and Twitter rather than trained educators. The risk isn’t that students won’t adopt AI; it’s that they’ll develop poor habits, miss critical thinking skills, or follow unreliable guidance. Schools that want to matter in AI education need to meet students where they are.

2

The Policy Confusion

Survey Data

One finding stands out: students don’t even know what their school’s AI policy is. The survey found that 31% of learners said their institution fully permits AI use, 11% said AI is banned entirely, and 13% said they simply don’t know their institution’s policy.

• Full permission: 31% say their school allows AI use
• Complete ban: 11% say their school prohibits AI entirely
• Don’t know: 13% are uncertain about their school’s policy
• Training improving: Institutional AI training is up—69% received training vs. 47% in 2024

The Honest Take

The fact that 13% of students don’t know their school’s AI policy suggests a communication problem. More concerning: 11% are at schools that ban AI entirely, putting those students at a disadvantage in learning tools they’ll need in the workforce. The middle ground—schools that allow AI with guidelines—is where most effective learning happens.

3

How Students Are Using AI

Survey Data

The survey reveals what students actually do with AI tools—and how that’s changed in a year. The top use has shifted from “enhancing learning and understanding” to “completing assignments more efficiently.”

• Completing assignments: 44% use AI to finish work more efficiently (top use in 2025)
• Enhancing understanding: 38% use AI to improve learning (was top use in 2024 at 45%)
• Exploring resources: 36% use AI to find additional learning materials
• Tutoring: 35% use AI for tutoring or assistance
• Training effectiveness: 35% said institutional training was “highly effective” vs. 15% in 2024

4

The Trust Problem

Analysis

Why do students trust social media more than schools? The survey doesn’t explicitly answer this, but the pattern suggests a failure of institutional responsiveness. Schools are often slower to adopt new tools, less transparent about policies, and more likely to issue blanket bans than nuanced guidance.

• Speed: Social media adapts to new AI tools instantly; schools take months or years to update policies
• Accessibility: YouTube tutorials are available 24/7; school training happens on schedules
• Relevance: Influencers demo the latest tools; school curricula may reference outdated systems
• Judgment: Social media is non-judgmental; schools may penalize AI use or treat it as cheating

The Honest Take

The JFF survey confirms what many educators suspected: students aren’t waiting for schools to figure out AI. They’re learning on their own, from sources that may or may not be accurate. Schools have an opportunity—if they can provide better guidance than a 15-second TikTok video. But that means being faster, clearer, and more practical than the algorithm.

What This Means for AI Education

The guidance gap is real. Schools are offering more training, but students aren’t seeking it out. This represents both a challenge and an opportunity for educators who want to be relevant.

Self-directed learning is the norm. Students are experimenting, watching YouTube, and figuring things out on their own. The question is whether they’re developing good habits or bad ones.

Policy clarity matters. The 13% who don’t know their school’s AI policy—and the 11% at schools that ban AI entirely—represent students who may be disadvantaged in developing AI literacy. Clear, permissive policies with guardrails seem to produce better outcomes than confusion or prohibition.

Share this article

𝕏 in ✉