Half of Students Are Using AI for Schoolwork. Only 1 in 3 Schools Has Written a Rule About It.

Half of Students Are Using AI for Schoolwork. Only 1 in 3 Schools Has Written a Rule About It.

A September 2025 study by Doss et al., published by the RAND Corporation, offers a national portrait of AI use across U.S. K–12 education. Drawing on data from over 15,000 students, teachers, parents, and school leaders, the findings show that while AI adoption has surged, institutional guidance has not kept pace.

By winter 2025, 54% of students and 53% of teachers were using AI for schoolwork or instruction — a 15- to 25-point increase in a single year. Yet fewer than 45% of principals reported having any AI policy, and only one-third of teachers said their school had clear academic integrity guidelines. Just 19% of students reported any teacher-led instruction on responsible AI use. Half of students feared being falsely accused of AI-related cheating — in a system that has yet to clearly define what that means.

The RAND team urges states and districts to develop coherent policies, provide tiered training for teachers and students, and extend AI literacy efforts to elementary schools.

Gobble's Take: Most students are using AI for school. Most schools still haven't told them what's allowed or what counts as cheating.

Source: Jace SoTL Substack

A Student Cloned a Teacher's Voice and Left Threatening Voicemails. There Was No Policy to Handle It.

A student in Texas used AI to clone a teacher's voice and left threatening voicemails. A student in New Jersey used deepfake technology to humiliate a classmate. In both cases, what failed wasn't just a rule — it was an entire framework for understanding what these tools can do and why that matters. These weren't elaborate schemes by bad actors; they were kids with access to powerful technology and no one who had taught them where the lines were.

That's the argument at the center of a widely-circulated newsletter aimed at school leaders: the most dangerous gap in K-12 AI governance isn't the absence of a written policy — it's the absence of a school culture that treats AI as something requiring judgment, not just compliance. Top-down bans have already proven they don't hold. What districts that navigated this more successfully built instead were governance structures with actual teeth: cross-departmental steering committees, legal compliance layers, and community trust built through transparency with parents. The key question, the author argues, isn't "do we have an AI policy?" but "does our school culture give students a framework for understanding what these tools can do before something goes wrong?"

For parents, the practical takeaway is worth asking at the next school board meeting or PTA night: does our school have an AI governance structure, or just a document? Who's on it, and when did they last update anything?

Gobble's Take: A policy nobody reads doesn't stop a student who doesn't know any better — school culture does.

Source: Miriam Bogler Substack

"Don't Plagiarize" Doesn't Cover AI. Schools Are Rewriting the Rules From Scratch.

For decades, academic integrity meant one thing: don't copy someone else's work. Now teachers are fielding questions their integrity policies were never designed to answer. Is it cheating to use AI to brainstorm? To outline? To draft and then edit? To translate a prompt into better English? Most existing school policies are silent on all of it — which means students and teachers are improvising.

Guidance from Victoria, Australia's Department of Education — one of the more developed frameworks currently in circulation — offers a window into what updated policies are starting to include. Schools are being asked to write explicit definitions of academic integrity, cheating, and plagiarism that account for AI, provide age-appropriate guidance on how to cite AI tools when their use is permitted, and spell out clear consequences for misuse. Critically, the guidance pushes schools to redesign assessments themselves — not just add a warning at the top. That means either restricting AI on high-stakes tests or reworking tasks so that students have to demonstrate skills AI can't fake: real-time in-class work, oral defenses, iterative drafts with teacher check-ins.

The underlying logic is that detection-first thinking — catch the cheater after the fact — doesn't work when the tool is this capable and this available. Prevention through better assessment design is the more durable solution.

Gobble's Take: If your school's academic integrity policy was written before 2023 and hasn't been touched since, it doesn't cover the tool your kid used last Tuesday.

Source: Victoria Department of Education

28 States Have Issued AI Guidance for Schools. Most of It Skips the Hard Part.

Twenty-eight U.S. states have now issued official guidance on AI use in K–12 classrooms, according to education researcher Carl Hendrick, urging districts to tackle privacy, bias, and appropriate use. The UK's Department for Education, alongside inspection bodies Ofsted and Ofqual, has done the same — framing AI around safeguarding, data protection, and academic integrity, and instructing inspectors to assess how schools use AI to support learning. Some early-adopting schools have gone further, building AI oversight committees that bring together teachers, administrators, parents, and students to audit algorithms and monitor usage.

But Hendrick's diagnosis of what's missing is pointed: most of what schools have produced are compliance-heavy documents. They cover what students can't do with AI, which vendors are FERPA- or GDPR-compliant, and what happens if someone gets caught. What they largely don't cover is what to actually do with AI — how to integrate it into curriculum, instruction, and assessment in ways grounded in how students actually learn. Hendrick describes AI-generated curricula appearing in some schools as "plastic plants": convincing on the surface, but lacking the depth and vitality that comes from expert teachers who understand their students. A checklist approach to AI policy, he argues, won't build that.

For school board members and administrators, the question worth asking isn't "do we have an AI policy?" but "does our policy say anything about curriculum, instruction, and assessment — or just compliance?"

Gobble's Take: Having an AI policy that covers data privacy but not learning design is like having a fire code but no fire drill.

Source: Carl Hendrick Substack

Quick Hits

• UAB Libraries are training Alabama's K-12 teachers on AI: The University of Alabama at Birmingham is sharing AI resources and skills with the state's K–12 educators, part of a growing wave of universities stepping in where district-level professional development has fallen short. The Birmingham Times
• Education Week asks whether AI teacher training is actually working: As schools rush to check the "PD" box, reporters are digging into whether the sessions teachers are sitting through translate into anything useful in the classroom. Education Week

In Case You Missed It

Yesterday's top stories:

• The Admin Team Went All-In on AI. Now the WiFi Is Down and Student Data Is at Risk.
• High Schoolers Are Faking AI Highlight Reels. Coaches Are Already Catching Them.