Teachers are drowning. The generative AI boom added a new tool to an already packed job, and according to MIT Tech Review, many educators still feel there’s no clear path forward. Organizations from OpenAI to UNESCO push AI use in the classroom, yet plenty of teachers remain confused about how to actually handle it.
One school offers a working model. MIT Tech Review profiled Cheshire Academy, a private boarding and day school in Connecticut with about 400 students in grades 9 through 12. Here’s how they approached it, step by step, and what you can borrow.
Quick Start
You’ll learn a practical framework for bringing AI into teaching without chaos: how to train staff, which tasks to hand to AI, and where to hold back. What you need: an open mind, a couple of AI tools to test, and a willingness to talk about AI with students even if you never touch it yourself.
Step 1: Don’t force it
Cheshire Academy administrators don’t require any instructor to use AI. Still, librarian and technology coordinator George Aiello says the “vast majority” use it in some way. Why it matters: adoption that grows by choice tends to stick. Mandates breed resistance and box-checking.
Step 2: Train on techniques, not specific tools
On the advice of consultants, the school trained staff on general methods for using AI rather than prescribing a particular product. The result is a patchwork: general chatbots like ChatGPT and Perplexity alongside specialized platforms like MagicSchool, built specifically for educators.
Why it matters: tools change fast. Teach people how to think about AI and they can adapt to whatever comes next, instead of being locked into one vendor.
Step 3: Teach the limits alongside the skills
The training covered how to craft useful prompts. It also stressed the technology’s limits, highlighting its potential for generating incorrect and biased responses.
This is the part most rollouts skip. If you teach prompting without teaching skepticism, you get confident nonsense passed off as fact. Cover both.
Step 4: Start with prep work
Teachers at Cheshire now often use generative AI to prepare class materials. In practice that means:
- Planning lessons
- Creating grading rubrics
- Building out supporting materials
Why start here: prep work is lower stakes. A weak lesson draft gets fixed before it reaches students. It’s the safest on-ramp to build confidence.
Step 5: Know where to hold back
Some teachers want to use AI to help give students feedback. None of them have done it yet. Concerns over quality, personalization, and privacy stopped them.
That restraint is the whole point. Feedback is personal, and student data is sensitive. Knowing where not to use AI is as important as knowing where to use it. Draw that line before you cross it.
Step 6: Address AI even if you don’t use it
Miriam Przybyla-Baum, who teaches French, doesn’t use AI herself. After nearly 30 years of teaching, she’s built up plenty of classroom materials. But she still deals with it, because she started seeing students use tools like Google Translate to take shortcuts years before ChatGPT launched.
The lesson: opting out of AI as a teacher doesn’t mean it’s absent from your classroom. Students bring it in regardless. You have to have a plan for that.
Why this matters
What stands out in the Cheshire approach is that it treats AI as a judgment problem, not a software problem. No mandate, no single tool, honest talk about limits, and clear boundaries around sensitive work. That’s a repeatable playbook, not a one-off.
Next steps
- Pick two AI tools and test them on one lesson plan and one rubric this week. Compare the output.
- Write down your own “do not use AI here” list before you expand usage. Privacy and student data come first.
- Fact-check every AI output against a source you trust. Treat it as a first draft, never a final answer.
- If you don’t plan to use AI yourself, still write a short classroom policy on how students may or may not use it.
Full details on Cheshire Academy’s approach are available at the original MIT Tech Review report.