8 Classroom AI Policies Developed by Teachers

At Brimmer and May, a pre-K to 12 private school in Massachusetts, it’s “mostly up to the individual teachers” to decide how and when AI should be used by students, says high school history teacher David Cutler. 

That’s a level of responsibility and autonomy he welcomes.

“We all teach different disciplines at different levels and we're the experts in the room,” Cutler surmises, reflecting on his school’s decision to give teachers discretion. “We're the ones best informed to make decisions, including those about AI and edtech."

But not every teacher feels well equipped to make that call. Confronted with a revolutionary new technology that's rapidly evolving, educators have little precedent to draw on: “We don't have a social contract,” for AI, explains Marc Watkins, a lecturer at the University of Mississippi and founder of the Mississippi AI Institute for Teachers. “We don't have an agreed-upon set of principles that we can all look at.” School and district leaders, too, often struggle to provide clarity, says veteran educator Leah Cleary: “This might be the messiest policy landscape I’ve ever tried to make sense of. We’ve got federal whiplash, state guidance that may or may not matter, districts scrambling to figure out if they even need a policy, and teachers like you and me, just trying to get through the day without completely redesigning our entire curriculum.” 

What is clear, however, is that student use of AI is outpacing policy. Over half of teens surveyed reported using chatbots to “search for information or get help with schoolwork,” according to a recent Pew Research Center report, but attitudes toward the technology are far from unanimous, says Victor Lee, faculty lead for AI and education at the Stanford Accelerator for Learning. Some students see AI “as a bad force in the world,” he explains, while others are caught between the obvious utility of AI tools and the fear of leaning too heavily on a technology that is quietly eroding their cognitive development.

Some educators are exploring new ways to meet the challenge. “I knew I needed to be clear with students so they aren't guessing what's allowed or not, but even more so because they don't know what responsible AI use looks like,” explains classroom educator Stacy Kratochvil. While it’s tempting to discuss policy once and move on, her goal is to sow the seeds for future conversations. “I recommend revisiting AI expectations before major assignments so students learn to make thoughtful decisions.”

We asked teachers across disciplines and grade levels to share the specific strategies they’ve developed to establish shared values, expectations, and policies about AI use in their classrooms.

STRUCTURED PERMISSION AND TRANSPARENCY

When AI tools first emerged, middle school science teacher Eric Cross chose not to ban them. He’d learn much more, he decided, from allowing students to use these tools openly and observing them. “It’s hard to learn about behavior when we push it into hiding,” he says. 

Disclosures Required: Transparency became the foundation of Cross’s policy: each assignment is clearly marked with the level of AI use (if any) he considers acceptable, and students must submit a standardized “AI Use Note” detailing their process. The template provides sentence starters that require students to document what AI tool they chose, the prompt they entered, and how this impacted their process. A student might submit the following, appended to a biology task:

I used Gemini to help me understand the word “homeostasis.” My prompt was, “Explain homeostasis for a 7th grade science student.” It helped me understand the idea. After that, I wrote my own sentence in my own words and added my own example about body temperature.

Ask Permission First: Borrowing from higher education, high school English teacher Brett Vogelsinger’s course syllabus takes a similar tack: “Like most major colleges, I require that you ask permission before inviting AI into your writing process so that we can together assess how it can be most helpful without taking away opportunities to learn the required skills for the course,” he writes. On major writing projects, students fill out a transparency survey, disclosing any AI or human assistance and inviting reflection. 

The Rules Apply to Everyone: When working with college students, Watkins begins by modeling transparency himself. His classroom policy delineates various tasks he will not use AI for—like providing feedback, grading work, and writing letters of recommendation—along with his reasoning: “These purposes… can impact the relationship I have with you, and that is something I value much more than efficiency,” he writes. 

Watkins also commits to disclosing his own AI use and requires students to do the same, inviting them to “craft their own statements” about how or whether they’ll use AI in his class. For clarity, Watkins labels each assignment with clear terms of engagement, like whether students can ask for help (and from whom), whether they can collaborate with others, and what level of AI use, if any, is permitted.

CO-CREATING SHARED VALUES

High school English teacher Cathleen Beachboard recalls that her first attempts at a classroom policy were little more than a list of AI dos and don’ts. “I now view the role of the policy as defining how learning should occur,” she says. 

Preserving Student Voice: Informed by research that suggests “humans naturally value things that have their fingerprints on them,” Beachboard opened the floor for conversation, posing questions like If AI is going to continue to be present in our lives, what type of classroom do we want? Students’ responses helped “develop values” instead of rules, which became the backbone of a formal policy that hangs on her classroom wall. 

Each year, as the technology and her students’ understanding of it evolve, Beachboard revisits the policy with her new classes, discussing what remains relevant and what can be added or challenged. “This collaborative process is extremely important because policies should not simply be handed down; they should be discussed, questioned, and improved upon via collaboration.” 

Co-Creating New Norms: Stacy Kratochvil takes a similar approach, prompting open class conversations first to introduce what AI is and help students consider its role in their lives. Next, students co-create an AI class agreement together. She starts by naming core values that the class agrees their potential AI use should support. For example: “Integrity—We’re honest about when and how we use AI.” On a worksheet, students suggest additional values, naming ways that AI might support or undermine them, then brainstorm AI use cases that should be allowed or prohibited, or that might require teacher insight. The final document is displayed prominently in her classroom and revisited throughout the year.

A Common Code of Ethics: College lecturer Sydney Sharkey, too, was craving more nuance in classroom AI conversation that often treats students as “cheaters-in-waiting.” So she “ditched the lecture about ‘AI misuse’” and opened a dialogue. 

With the assistance of guiding questions like What types of AI use feel ethical, helpful, or appropriate in a writing course?, Sharkey asked students to document their thoughts on a shared Padlet board and reply to at least two of their classmates. “Several students admitted they’d experimented with AI before but hadn’t felt comfortable asking what ‘counted’ as cheating,” Sharkey explains. “Others expressed relief that we were even having this conversation.” She then condensed and synthesized the discussion into a co-created classroom policy, including approved and inappropriate uses of AI, and an expectation that students would transparently disclose their usage.

DISCOURAGING OR HEAVILY RESTRICTING AI USE

Some educators believe the use of AI at any stage of learning may be damaging. 

In his history classroom, Cutler sees his role as helping students to become competent, critical thinkers and writers. “Anything that is able to impede that I want to limit to the best of my ability,” he explains. For this reason, he keeps generative AI use to a minimum. “I really prefer that they don't use it,” and if they do, ”I want it to be under my supervision.”

In Your Own Style: Cutler’s school’s broader Artificial Intelligence and Academic Work Policy clearly states that “while an AI tool is helpful in generating ideas,” students “who take the ideas or works of others” to pass off as their own are “considered to be plagiarizing” and will earn a zero. This includes using AI software like Quillbot or Grammarly Premium to enhance the language, tone, or sophistication of their writing; asking AI chatbots to write full sentences or paragraphs of text; or sending AI off to conduct research on their behalf uncited. 

With this high-level foundation set, his classroom policy doesn’t mince words:

Students are strictly prohibited from using Al to assist in any production of work, unless otherwise noted by Mr. Cutler. Use of Al will result in an automatic "zero," without the ability to revise or redo for any credit.

A Defense of Humanism: In contrast, students “have noted that they’ve never seen an AI policy this detailed,” high school humanities teacher Alexa Garvoille says of her syllabus. Across several paragraphs, Garvoille elaborates on why she would prefer students “not use ChatGPT, Gemini, Claude, AI-based editors, or other LLMs to generate ideas, summarize readings, or brainstorm for this class.” 

Acknowledging that it’s simply too challenging to enforce an all-out ban, like Cross and Vogelsinger, she requires students to transparently describe how they’ve used AI on each assignment. The length of her policy, she admits, “didn’t stop a few students from using AI,” but interventions were “conversation-based and focused on redoing work.” 

Sign on the Dotted Line: To affirm they’ve understood his policy and keep them accountable, Cutler asks each student to sign it. “There were a few instances of clear AI use,” he says. A couple of students claimed they weren’t aware of the AI policy, which Cutler quickly dismissed: “I've said it a million times in class each day for the last six weeks. You signed your name to a statement that you understand this. And ‘no AI’ is also all over the assignments.” Despite his policy, he allowed those students to redo the work for up to half credit, an offer some (not all) took him up on: “I try to be understanding when mistakes are made where there's no malicious intent behind them.”