60% of Teachers Say AI Is Eroding Student Trust

An NPR/Ipsos poll of 545 K-12 teachers, published June 5, 2026, found that nearly 60% believe AI is eroding trust between them and their students. 55% say students mostly use AI as a shortcut, and 54% say it's making critical thinking harder to build. The response from 40% of teachers has been to assign more handwritten and in-class work - reverting to methods that predate the problem instead of solving it.

60% of Teachers Say AI Is Eroding Trust With Their Students. That's the Real Cost, Not the Cheating.

Most coverage of AI in classrooms leads with cheating. The more precise finding, from a nationally representative NPR/Ipsos poll of 545 K-12 teachers published June 5, 2026, is about something harder to repair than a grade dispute.

Nearly 60% of teachers say AI is eroding trust between them and their students.

Not "AI is being misused." Trust - the thing a classroom runs on - is degrading.

The numbers behind the headline

• 55% of teachers say students mostly use AI as a shortcut to avoid doing more work.
• 54% say AI is making it harder for students to build critical thinking skills.
• Only 35% of schools that provide AI software have a formal policy on it.
• 60% of teachers report their school offers no guidance on AI, or they're unsure if it does.
• Despite all this, 60% of teachers use AI themselves for work tasks, and nearly 75% believe AI's overall impact on education will exceed that of the internet or computers.

The picture isn't teachers rejecting AI. It's teachers using it themselves while distrusting how their students use it, with almost no institutional guidance bridging the two.

Why "more handwritten work" is a symptom, not a fix

40% of teachers say they now require more handwritten assignments, and 40% require more in-class work, specifically because of AI. That's a rational short-term response to an unsolved problem: if you can't verify what a student did at home, move the work somewhere you can watch it happen.

But this reverts the classroom to methods that predate the actual insight AI-era assessment should be capable of. In-class handwriting proves a student wrote something in front of you. It doesn't diagnose what they misunderstood, and it doesn't scale - a teacher watching 28 students write by hand still can't tell, in real time, which student's error is a careless mistake and which is a structural gap in fraction division.

Where the trust actually breaks - and where it doesn't have to

Trust erodes at a specific point: when the teacher can no longer treat a piece of student work as evidence of what the student knows. Once that's gone, everything downstream - grading, parent conferences, intervention decisions - is built on sand.

The fix isn't detection software. Detection tools carry their own well-documented reliability problems (Ohio's education department has flagged AI-detection false-positive rates as high as 61.3% for non-native English speakers' writing, a serious equity issue on its own). Chasing detection accuracy treats the symptom.

The fix is separating two different jobs that got collapsed into one: AI that a student can use to produce output, and AI that helps a teacher understand what actually happened during learning. When the first exists in a classroom, the second becomes necessary to restore what was lost. A system that runs on the assessment evidence a school already collects - never interacting with the student, only diagnosing what the evidence shows - doesn't need to detect anything, because nothing was hidden from it in the first place.

That is a different kind of trust: not "we caught you," but "we can see clearly, because the student never had a tool sitting between us and their actual work."

See Teacher-Facing AI in Action.

Related reading

• why teachers use AI but distrust student use
• AI detector false positives
• Ohio AI policy deadline
• cognitive debt and student work
• teacher-facing AI

FAQ

What percentage of teachers think AI is eroding trust with students?

Nearly 60%, according to an NPR/Ipsos poll of 545 K-12 teachers published June 5, 2026.

Do teachers think students use AI mainly to cheat, or for other reasons too?

55% of teachers say students mostly use AI as a shortcut to avoid doing more work - not necessarily framed as cheating, but as avoidance.

How are teachers responding to declining trust caused by AI?

40% now require more handwritten work, and 40% require more in-class work, specifically in response to AI use. Both are reactive measures rather than diagnostic ones.

Do most schools have formal AI policies for teachers or students?

No. Only 35% of schools that provide AI software have a formal policy governing its use, and 60% of teachers report no guidance or are unsure whether any exists.

Sources

• NPR, "Poll: Teachers worry AI is impacting students' critical thinking," June 5, 2026
• KJK, Ohio AI policy analysis (61.3% false positive rate)