Lesson Plan With AI
AI can save time on lesson planning, but it cannot decide what your students need to learn or whether a lesson will work in your room. Treat it as a drafting assistant. You set the standard, learning goal, evidence of learning, time, materials, and student needs. Then you use AI to generate options, review the output, and make the final calls.
Table of Contents
That distinction matters. In a 2025 Gallup survey of 2,232 U.S. public K-12 teachers, 60% reported using an AI tool for work. Teachers who used AI at least weekly estimated that it saved them 5.9 hours per week, but those are self-reported time estimates, not proof that every AI-generated lesson is effective (Gallup).
Key Takeaways
- Start with a verified standard, an observable objective, and evidence of learning before opening an AI tool.
- Give AI enough classroom context to draft a useful plan, then revise one dimension at a time.
- Check accuracy, rigor, accessibility, pacing, bias, privacy, and school policy before teaching the lesson.
- The best use of AI is to create a stronger first draft so you have more time for instructional judgment.
Use AI as a planning assistant, not the teacher
AI is good at producing a first-pass sequence, generating activity options, adapting instructions, and suggesting assessment questions. It is not reliable enough to determine the right standard, invent a factually correct answer key, or know the specific dynamics of your class without your review.
UNESCO’s teacher AI framework puts human agency, ethics, AI foundations, pedagogy, and professional learning at the center of teacher competence. The framework describes 15 competencies across those five dimensions, which is a useful reminder that responsible use is broader than writing a good prompt (UNESCO).
Keep two uses separate:
- Teacher-only planning: You use AI to draft or revise a lesson before class.
- Student-facing AI: Students use AI during an activity or assignment.
This guide focuses on teacher-only planning. If students will use AI, plan separate instructions for disclosure, access, academic integrity, source checking, and alternatives for students who cannot or should not use the tool.
Step 1: Gather the inputs before you prompt

The quality of the first draft depends on the quality of the information you provide. Do not begin with “make a fun lesson on photosynthesis.” That produces a generic response because the AI has to guess the goal, prior knowledge, lesson length, materials, and level of rigor.
Before prompting, write down these nine inputs:
- Course, grade, and subject
- The exact, verified standard text and its official source
- Topic or concept
- What students already know
- Likely misconceptions
- An observable learning objective
- Evidence that will show learning
- Time, room, materials, technology, and other constraints
- General learning needs, described without personally identifiable student information
Use the official standard text rather than asking AI to supply a standard code. AI may return a plausible-looking code or a partial description. Paste the standard you have already checked, then ask the tool to show how its suggested activities support that standard.
For learner needs, use general descriptions such as “some students benefit from vocabulary support and visual sequencing” or “several students are ready for an extension that requires application.” Do not paste names, grades attached to names, IEP details, behavior records, health information, or other personally identifiable data into an unapproved tool.
Step 2: Define the objective and the evidence of learning
Choose what students should be able to do before you choose the activity. A lively activity can still miss the learning goal if you do not know what evidence you expect to see.
For example, this objective is too broad:
Students will understand mitosis.
This version is observable:
Students will sequence the stages of mitosis and explain how mitosis supports growth and tissue repair using a labeled model and a four-sentence explanation.
The second objective tells you what the lesson needs to produce. Students need to sequence, label, and explain. Your exit ticket should ask them to do one or more of those things. If the exit ticket only asks students to name stages from memory, it does not fully test the stated objective.
AI can help you propose stronger wording, examples, or checks for understanding. It should not replace the teacher’s decision about the appropriate intellectual demand. Ask yourself: what will a student say, write, build, solve, or defend if they have met the goal?
Step 3: Give AI a structured lesson-planning prompt
A structured prompt makes it easier to inspect the answer. It also makes it harder for the tool to hide its assumptions behind polished prose. Copy this prompt and replace the bracketed fields.
Act as an instructional planning assistant. Draft a lesson for teacher review.
Do not invent standards, source claims, student details, or school policies.
Mark any uncertainty clearly.
Course and grade: [COURSE AND GRADE]
Topic: [TOPIC]
Verified standard text: [PASTE OFFICIAL STANDARD TEXT]
Prior knowledge: [WHAT STUDENTS ALREADY KNOW]
Likely misconceptions: [MISCONCEPTIONS]
Learning objective: [OBSERVABLE OBJECTIVE]
Success criteria: [EVIDENCE OF LEARNING]
Duration: [MINUTES]
Class context: [NON-IDENTIFYING DESCRIPTION OF LEARNER NEEDS]
Materials and technology: [AVAILABLE RESOURCES]
Constraints: [ROOM, GROUP SIZE, POLICY, ACCESSIBILITY, OR OTHER LIMITS]
Create a timed lesson plan with:
1. A short retrieval warm-up
2. Teacher modeling or explanation
3. Guided practice
4. An active student task
5. Checks for understanding across a broad sample of students
6. Differentiation and accessibility options
7. An exit ticket aligned to the objective
8. Anticipated misconceptions and teacher responses
9. Materials and preparation
10. A timing total that matches the stated duration
For every activity, state how it supports the learning objective. End with
“Teacher verification required” and list the standards, facts, resources,
assumptions, and privacy or safety issues I should check.This prompt asks AI to produce a draft that is easy to review. It does not ask the tool to make teaching decisions in your place.
Step 4: Improve the draft one decision at a time
Do not try to repair every flaw with one large follow-up prompt. Review a single planning dimension at a time so you can see exactly what changed.
Try this sequence after the first draft:
- “Audit the timing. Show the total and cut low-value transitions.”
- “Check whether every activity produces evidence for the objective. Remove or revise anything that does not.”
- “Increase the cognitive demand without increasing the reading load.”
- “Add one scaffold and one extension that preserve the same core objective.”
- “List factual claims, answer keys, examples, and resources that require an authoritative check.”
This kind of revision matches how many teachers already use AI. Gallup found that teachers most commonly reported monthly AI use for preparing to teach, 37%, making worksheets or activities, 33%, and modifying materials to meet student needs, 28% (Gallup). Use those strengths to generate alternatives, not to bypass professional review.
Step 5: Differentiate the lesson without exposing student data
AI can create useful variations when you describe the learning barrier and keep the core goal intact. For example, ask for a vocabulary support, visual sequence cards, worked examples, sentence frames, an oral response option, or a richer application task for students who are ready to extend their learning.
Universal Design for Learning is a helpful way to organize those options. CAST groups its guidelines around engagement, representation, and action and expression (CAST). In practical terms, that means asking how students can connect to the task, access the information, and demonstrate their learning in more than one appropriate way.
Keep privacy at the front of the process. The U.S. Department of Education advises teachers to check with school or district administration about whether an online service is approved before using it in a course. Its guidance also points teachers toward IT staff for privacy and security questions (U.S. Department of Education).
That is why your prompt should use general classroom descriptions. “Provide supports for students who need vocabulary scaffolds” is useful. “Rewrite this plan for Maya, who has these records and grades” is not an appropriate prompt for a general AI tool.
Step 6: Approve the lesson with the ALIGN checklist

Before you teach an AI-assisted plan, run this quick review. It turns a vague instruction to “edit the output” into a set of decisions you can actually make.
A: Aim
Does the objective match the verified standard and the place of this lesson in the course sequence? Check that the plan does not claim a standard alignment that you have not verified yourself.
L: Learning evidence
Do the student tasks and exit ticket reveal the target learning? A worksheet, discussion, or game belongs in the plan only if it gives you usable evidence connected to the objective.
I: Integrity
Are facts, calculations, quotations, links, examples, images, and answer keys correct? Check primary or authoritative sources for factual material. AI can state false information confidently, especially when it creates subject examples or source lists.
G: Group access
Can your students access the instructions, materials, and task? Check reading load, vocabulary, cultural assumptions, available technology, accessibility supports, and extension opportunities. Do not lower the core goal by default. Change the path students use to reach it when that is appropriate.
N: Nonnegotiables
Does the plan fit the available time, materials, safety requirements, privacy rules, school policy, and your own professional judgment? The U.S. Department of Education notes that AI can automate pattern detection and instructional decisions, which can introduce bias and unfairness if schools do not govern its use (U.S. Department of Education). Your review is the safeguard.
You can use a second AI prompt as a red-team aid: “Act as a skeptical department chair. Identify alignment, rigor, timing, accessibility, factual, bias, safety, and privacy failures. Do not rewrite the lesson.” Read its answer as a list of possibilities to check, not as a final judgment.
Worked example: a Grade 10 biology lesson on mitosis

Imagine you have 55 minutes with a Grade 10 biology class. Students have already learned that cells contain chromosomes, but they often confuse mitosis with meiosis. You want students to sequence mitosis and explain why it matters for growth and repair.
The weak prompt is simple:
Make a fun high school lesson on mitosis.
It gives AI almost no useful planning information. A stronger version includes the verified local standard text, the objective above, a 55-minute limit, available sequence cards and microscopes, the likely confusion with meiosis, and the requirement for an exit ticket.
After the AI drafts a plan, your approved version might look like this:
| Time | Lesson segment | Evidence of learning |
|---|---|---|
| 5 minutes | Retrieval warm-up: students sketch or label a chromosome and name one function. | Quick scan of prior knowledge. |
| 8 minutes | Model the relationship between chromosomes, DNA replication, and mitosis using a visual sequence. | Cold-call and partner explanation checks. |
| 12 minutes | Guided practice: pairs order stage cards and justify each placement. | Students explain the evidence for their order. |
| 15 minutes | Application task: students create a labeled model or annotated sequence, then connect mitosis to growth or repair. | Model and written explanation meet the objective. |
| 8 minutes | Check for understanding: compare one feature of mitosis and meiosis to surface the likely misconception. | Individual responses show whether the distinction is secure. |
| 5 minutes | Exit ticket: sequence two stages and explain in four sentences why mitosis matters for a body process. | Direct evidence aligned with the objective. |
| 2 minutes | Buffer and collection. | Realistic pacing. |
Notice the change from “fun” to evidence. A game can remain part of the lesson, but it is not the reason the lesson is sound. The sequence-card activity stays because students must justify their reasoning. If it were only a race to arrange cards, it would need revision.
After class, record what ran long, which misconception remained, and what the exit tickets showed. You can give AI a de-identified summary, ask for two revisions, and choose the one that fits your next class. That final feedback loop is where a generic draft becomes a better lesson over time.
Frequently asked questions
Can AI create a complete lesson plan?
It can create a complete-looking draft. A classroom-ready plan still needs teacher approval because AI does not reliably know your local curriculum, students, equipment, school policy, or the factual accuracy of every example it generates.
Which AI tool is best for lesson planning?
There is no single best tool for every school. Start with district approval, privacy terms, accessibility, source handling, output controls, export options, and cost. Product features and policies change, so verify the current terms before adopting a tool.
Can I put student information into an AI tool?
Follow your school or district policy and use only approved services. For general planning, do not provide personally identifiable student information to a general AI tool. Describe learning needs in broad, non-identifying terms instead.
How long should I spend reviewing an AI-generated lesson?
There is no universal benchmark. Review takes longer when a lesson includes new content, high-stakes assessment, external sources, safety concerns, or complex access needs. The right question is whether you have checked everything that could affect student learning or safety.
Plan, prompt, prove
AI can remove some of the blank-page work from lesson planning. Use that time to improve the parts that matter most: a clear objective, meaningful evidence of learning, accurate content, accessible materials, and responsive teaching.
Plan the lesson before prompting. Prompt AI with the constraints it needs. Then prove the draft is appropriate for your students before it reaches the classroom.
