What 20 U.S.C. § 1232g actually says, how to read your district's AI policy in fifteen minutes, and how to walk into your principal's office prepared rather than apologetic.
Leave with a working understanding of FERPA's core rule (20 U.S.C. § 1232g) as it applies to AI tools, a printable checklist for reading your district's AI policy, and a five-minute conversation script for telling your principal what you are doing and why it is compliant.
The Family Educational Rights and Privacy Act, codified at 20 U.S.C. § 1232g, conditions federal education funding on a school's compliance with rules about student "education records." The operative language teachers need to know is the prohibition on disclosure without consent.
The statute states that funds "shall not be made available" to any school "which has a policy or practice of permitting the release of education records (or personally identifiable information contained therein . . .) of students without the written consent of their parents to any individual, agency, or organization" except under specific listed exceptions. (20 U.S.C. § 1232g(b)(1)).
The phrase to circle is "personally identifiable information contained therein." Even if you do not send the whole record, sending pieces of it that identify the student is the same problem. A name plus a grade is PII. A name plus an IEP accommodation is PII. A name plus a behavior incident is PII. Pasting any of those into a third party that has not signed a FERPA-compliant data agreement with your district is a disclosure event.
FERPA permits disclosure to "other school officials, including teachers, within the agency or institution whom the agency or institution has determined to have legitimate educational interests" (20 U.S.C. § 1232g(b)(1)(A)). The federal regulations at 34 C.F.R. § 99.31(a)(1)(i)(B) extend this to certain contractors, consultants, and vendors who perform a service the school would otherwise use its own employees for — provided the vendor is under the direct control of the school and is subject to FERPA's use-and-redisclosure rules.
This is the doorway through which districts authorize specific enterprise AI tools. When a district signs a data-processing agreement with Microsoft Copilot for Education, Google Gemini for Education, OpenAI's enterprise/edu offering, or Anthropic's enterprise offering, the agreement is what makes the AI tool a "school official" for FERPA purposes. The free consumer versions of those same tools are typically not covered by such agreements and therefore do not benefit from this exception.
Pull up your district's AI policy and answer these twelve questions in order. The whole exercise should take fifteen minutes. If your district does not have an AI policy yet, the same checklist tells you what to ask the people who are writing one.
Print your answers. Put them in your binder. The most important professional development you will do this year on AI is reading these twelve answers from your own district.
Most district AI policies, in 2026, are still incomplete. They were written to react to the first wave of student-cheating concerns and they often omit the teacher-workflow questions this course covered. If your policy is silent on a workflow you want to use — for example, anonymized rubric-aligned essay feedback — the safest path is to ask, in writing, before adopting the workflow.
"Silent" does not mean "permitted." It means "not yet decided." Your principal or curriculum director needs the question on their desk so they can answer it for you and for every other teacher considering the same workflow. The question itself is a contribution to the policy — you are doing the district a favor by surfacing it.
Walk in with a one-page summary, not a presentation. The script below is the version I have seen work in three different districts. Adapt the language to your principal and your school culture.
"I wanted to give you a quick heads-up on how I'm using AI in my
prep work this year, and ask one or two questions about our policy.
What I'm using it for: drafting lesson plans, drafting feedback on
student writing against my rubrics, drafting parent newsletters,
and adapting assignments for the accommodations my IEP-served
students need.
What I'm NOT putting into any AI tool: student names, IEP content,
behavior records, individual grades, or anything that identifies
a specific student. I use placeholder tokens like [STUDENT] and
fill in the names offline.
The rules I'm following: FERPA at 20 USC 1232g, our district's
AI policy [cite section if you have it], and the rule that the
grade decision is always mine.
Where I'd like your input: [pick ONE specific question - e.g.,
'is our district's enterprise tool the right one for this, or
should I be using something else?']
I have a one-pager with my prompts and what I'm using them for.
Want me to leave it with you?"That conversation accomplishes three things. It tells your principal you are being thoughtful. It puts your specific workflow on the record before anyone else asks. And it positions you as a teacher who is building professional capacity, not a teacher cutting corners. Principals universally prefer the first to the second.
FERPA does not prohibit AI use in schools. It does not prohibit teachers from using AI for prep work. It does not prohibit AI-drafted feedback. What it does is restrict disclosure of personally identifiable information from education records. The workflows in this course are designed around that restriction. They use AI for the structural drafting that does not require PII, and they keep the PII offline where it belongs.
The headline-grabbing "FERPA bans ChatGPT" stories that show up in education media are usually about districts banning specific consumer tools because of disclosure risk, not about FERPA itself banning AI. The distinction matters because it tells you what to fix. If your workflow has zero PII flowing to a third party, you have zero FERPA problem regardless of which tool you are using.
You have spent five days building a stack of prompts and a stack of policy understanding. The point was always to use them. Pick the single workflow from this course that buys back the most of your evening — for most teachers it is either lesson planning (Day 1) or rubric feedback (Day 2) — and use it for the next five school days. At the end of the week, count the hours saved and the hours misspent, refine the prompt, and add the next workflow. Compounding starts when you stop reading and start using.
Thank you for spending this week with me. The work you do matters; if AI gives you back a few hours a week to do more of it, then this course did its job.