AI for Teachers 2026: Lesson Plans, Grading, FERPA

Teaching is a vocation that asks more of a person than the day will hold. A typical K-12 teacher prepares lessons for five subjects across multiple ability levels, grades dozens of papers a week, communicates with parents, supports students with IEPs and 504 plans, and tries to leave the building before dark. The arrival of practical AI in the classroom is the first development in a decade that meaningfully redistributes that load. Used well, AI gives back planning hours and grading hours that you can return to your students, your family, and your sleep. Used carelessly, it puts student records, your contract, and your district in front of regulators.

This guide is the condensed version of our five-day course, AI for K-12 Teachers. It walks through where AI is helping teachers right now, the prompts that work, and the rules - especially FERPA - that have to govern every workflow. By the end you will know what to start using on Monday morning, what never to paste into a public tool, and how to bring a clear AI policy to your classroom and your principal.

Why AI matters for teachers in 2026

The Department of Education's 2024 and 2025 reports on AI in education describe a profession that is curious, cautious, and overdue for time relief. RAND Corporation surveys found that planning and grading consume the largest share of teacher non-instructional time. Most teachers want help; few have a clear policy from their district.

Three concrete examples this year. First, a fifth-grade teacher in Ohio uses AI to draft three differentiated reading-comprehension worksheets for a Charlotte's Web unit, one for grade-level readers, one for English-learner students, one for advanced readers, in twenty minutes total. Second, a high-school chemistry teacher in California uses AI to draft rubric-aligned feedback on twenty-five lab reports, edits, and saves four hours of grading. Third, a middle-school principal in Texas pilots a school-wide AI policy that lets teachers use approved tools with FERPA-protected data and bans unapproved consumer tools entirely. None of these educators handed teaching to a machine. Every one of them got hours back.

The five things every teacher needs to know about AI

1. AI for lesson planning and unit design

Lesson planning is the highest-leverage starting point. Feed AI a standard, a grade level, and a duration; ask for a lesson plan with objectives, materials, sequence, and assessment; review and adapt. AI is excellent at the structural work and weak at knowing your students - which is why you finish the plan, not the AI.

Write a 45-minute fifth-grade math lesson on equivalent fractions
aligned to CCSS 5.NF.A.1. Include: objective, warm-up, mini-lesson,
guided practice, independent practice, exit ticket, and a quick
formative-check question.

The pitfall: never paste student names or identifying details into a consumer AI tool. Use placeholders. See Day 1 for the planning workflow.

2. AI for differentiation and IEP-aligned materials

Differentiation is a beautiful use case. AI can take one core text and produce three reading levels, four scaffolds, two ELL supports, and one extension within minutes. The teacher curates and adapts.

From this 600-word reading on photosynthesis, produce three versions:
A) sixth-grade Lexile, B) fourth-grade Lexile, C) ELL-supported with key vocab
in bold and a glossary at the end. Keep the same key concepts in all three.

The pitfall: IEP-related accommodations are protected records. Never paste a student's IEP into a non-approved tool. Use abstracted prompts ("a student with a reading-fluency goal") instead of names.

3. AI for grading, feedback, and rubrics

AI is excellent at producing rubric-aligned feedback drafts on student writing. The teacher reads each draft, edits, and signs the grade. This combination preserves teacher judgment while removing the typing.

Using this 4-point rubric (Idea, Organization, Voice, Conventions),
produce written feedback on the attached student essay. Keep tone
warm and growth-oriented. Suggest one specific revision for each rubric category.
Do not assign a final score - I will.

The pitfall: automated grades without teacher review breach professional standards in most districts. The teacher signs.

4. AI for parent and family communications

Parent communications eat hours and are emotionally costly when a hard message is needed. AI helps draft a calm, structured email; the teacher edits to add specifics and warmth. Translate into the family's home language with care; review the translation if you can.

Write a calm parent email about a student who has been turning in
incomplete homework for two weeks. Tone: supportive, factual, partnership-focused.
Include one specific next step we can try together. Under 150 words.

The pitfall: if you paste a child's name and behavior detail into a public AI tool, you may be disclosing an education record under FERPA. Strip names; fill them in locally before sending.

5. Student AI use and classroom policy

Students arrive with AI in their pockets. A clear, written classroom policy beats a quiet ban every time. Define when AI is allowed (brainstorming, study help), when it is not (final drafts, assessments), and how to disclose use. Tie the policy to the school honor code.

Draft a one-page classroom AI policy for a sophomore English class.
Cover: when AI is allowed, when it is not, how to disclose use,
and consequences for misuse. Sixth-grade reading level so families understand.

The pitfall: COPPA, 15 USC 6501-6506, requires verifiable parental consent for AI tools that collect personal information from students under 13. Older students still owe truthful disclosure of AI use.

The FERPA compliance summary

• FERPA, 20 USC 1232g and 34 CFR Part 99, protects personally identifiable information in education records. Disclosure to a non-approved AI tool can be a FERPA violation.
• 34 CFR 99.31(a)(1) school official exception permits disclosure to a vendor only if the vendor is under direct control of the school and uses the data only for the school's authorized purpose.
• COPPA, 15 USC 6501-6506, requires verifiable parental consent for under-13 students using tools that collect personal information.
• IDEA, 20 USC 1400, protects IEP records as education records and requires their confidential treatment.
• State student data privacy laws (CA SOPIPA, IL SOPPA, CO HB22-1108) layer additional vendor obligations.
• District AUPs and AI policies are the operational layer. If your district has a written AI policy, that policy is the floor.

What to do this week

1. Find your district's approved AI tool list. If none exists, ask your principal in writing.
2. Pick one upcoming lesson and run an AI lesson-plan draft using only the standard and grade level - no student names.
3. Build one differentiation set (three reading levels) for an existing text.
4. Draft three rubric-aligned feedback comments on a small writing sample. Edit. Sign.
5. Write a one-page classroom AI policy for your students' families. Send for principal review.

FAQ

Can I paste student names into ChatGPT?

Not the consumer version. FERPA, 20 USC 1232g, requires that any vendor receiving education records be under direct control of the district. Strip names or use a district-approved tool.

Will AI replace teachers?

No. AI accelerates planning, materials, grading drafts, and family communications. Relationship, judgment, and teaching stay with the educator.

Can AI grade student writing?

AI can produce a rubric-aligned draft of feedback. The teacher reviews and owns the final grade.

What about students using AI?

COPPA requires verifiable parental consent for students under 13 using AI tools that collect personal information. Establish a written classroom policy.

Can I send AI-drafted parent emails?

Yes, with the same care as any communication. Avoid pasting student identifiers into a non-approved tool. Use placeholders, then fill in names locally.