AI for Teachers: Lesson Plans, Grading, and Sanity

The Teacher's Impossible Math

You teach 120 students. Each needs personalized feedback. Each assignment takes 5 minutes to grade thoughtfully. That's 600 minutes — 10 hours — per assignment. You assign one per week. That's 10 hours of grading on top of 40 hours of teaching, planning, meetings, and being a human being who presumably wants to eat dinner and see their family.

The math doesn't work. It has never worked. Teachers have been doing it anyway through a combination of heroics, guilt, and accepting that "good enough" feedback is the best they can offer.

AI doesn't fix the structural problem of underfunded, understaffed education. But it can give you back your evenings. And that's not nothing.

Lesson Planning: From Hours to Minutes

The Starting Point

Traditional lesson planning: stare at curriculum standards, find or create materials, differentiate for multiple levels, align to assessments, format for the district template. Two to four hours for one solid lesson.

AI-assisted lesson planning: describe what you're teaching, who you're teaching, and what outcome you want. Get a complete lesson plan in five minutes. Spend 20 minutes customizing it for your specific class.

That's not a marginal improvement. That's a category change.

The Prompt That Works

"I'm teaching [subject] to [grade level]. The standard is [specific standard]. My class has [description of learners — reading levels, English learners, gifted students, IEP needs]. I have [time] minutes. I need a lesson that includes direct instruction, guided practice, independent practice, and an exit ticket. Include differentiation for three levels."

The output isn't perfect — it never is — but it's a solid 80% draft that captures the pedagogical structure, the time allocation, and the differentiation you specified. Your expertise fills in the remaining 20%: the specific examples your students will connect with, the management strategies for your particular class, the inside jokes and cultural references that make a lesson feel alive.

Building a Lesson Library

Over time, your prompts become a library. "The AP US History lesson prompt." "The 7th grade ELA argumentative writing prompt." "The physics lab report scaffold." Each one refined over multiple uses, incorporating student feedback, assessment data, and your evolving understanding of what works.

The 📜History Timeline prompt is particularly useful for history and social studies teachers. It generates interactive timelines that turn static curriculum into explorable narratives. Students engage differently when they can ask "but what happened next?" and get answers.

Grading: The Game-Changer

This is the use case teachers are most excited — and most nervous — about. Let me be clear about what works and what doesn't.

What AI Does Well

First-pass feedback on writing. Give AI a rubric and a student essay, and it will generate detailed feedback aligned to your criteria. Grammar, organization, evidence use, argument strength — it catches everything.

Consistency. By essay 87, your human feedback is objectively worse than your feedback on essay 3. You're tired, you're rushed, and the quality suffers. AI gives essay 87 the same attention as essay 3.

Rubric-aligned scoring. If you have a clear rubric, AI can suggest scores with explanations. "This essay earns a 3/4 on evidence use because it includes two specific examples but doesn't explain how they support the thesis."

What AI Does Not Do Well

Replace your professional judgment. AI scores are suggestions, not grades. You review, adjust, and override. Your knowledge of each student — their growth, their challenges, their effort — matters in ways that AI cannot capture.

Handle creative or unconventional work. A student who writes a brilliant essay in a format that defies your rubric will get flagged as non-conforming. Your job is to recognize that brilliance.

Detect the subtle stuff. A student whose writing suddenly gets dramatically better might have gotten help from AI themselves. A student whose writing gets dramatically worse might be going through something. AI catches neither of these.

The Grading Workflow

1. AI first pass. Upload the batch. AI generates feedback and suggested scores for each.
2. Your review. Read AI feedback alongside student work. Agree, disagree, modify. This takes about 2 minutes per paper instead of 5.
3. Personal notes. Add the things only you can: "I can see you worked really hard on your introduction — it's much stronger than your last essay." AI can't write that. You can.
4. Final scores. You make the final call. Always.

This workflow cuts grading time by roughly 50% while actually improving feedback quality because AI catches things you'd miss at hour nine of grading, and you add the human elements that AI misses entirely.

Differentiation: Finally Practical

Every teacher knows differentiation is best practice. Every teacher also knows it's nearly impossible at scale. Creating three versions of every activity for three proficiency levels, times four preps, is a full-time job on top of a full-time job.

AI makes it practical:

"Take this reading passage and create three versions: one at grade level, one simplified for students reading two years below grade level (maintain the key concepts but simplify vocabulary and sentence structure), and one extended version for advanced students (add analytical questions and primary source connections)."

Three versions. Two minutes. What used to be Saturday's project is now part of the planning process.

The 💪Workout Generator prompt approach applies here too — just as fitness plans should adapt to individual capacity, lesson materials should adapt to individual learning levels. The principle is the same; only the domain changes.

Parent Communication

Writing report card comments, email responses, and conference notes is time-consuming and emotionally draining. AI helps with the mechanical parts:

"Write a parent email about [student name]'s progress in math. Strengths: strong mental math skills, participates actively in group work. Areas for growth: showing work on paper, word problem comprehension. Tone: warm, encouraging, specific. Include one suggestion for home practice."

Two minutes instead of fifteen. Multiply by 120 students and you've saved entire days.

The Ethics Conversation

Let's address this directly: some people think AI in grading is lazy or unethical. Here's why I disagree.

The current system is the ethical problem. When teachers are so overwhelmed that they give superficial feedback, grade inconsistently because of fatigue, or simply can't differentiate effectively — that's the injustice. Students at the end of the grading stack get worse feedback than students at the beginning. Students in larger classes get less personal attention than students in smaller ones.

AI helps equalize. When every student's essay gets thorough feedback — partly from AI, refined by a human — the quality floor rises for everyone.

Transparency matters. I tell my students I use AI as a grading assistant, the same way I might use a teaching assistant. The final grades and personal feedback are mine. The rubric-aligned analysis is AI-assisted. Students understand and appreciate the faster turnaround.

The time you save is the point. Every hour saved on grading is an hour you can spend planning better lessons, meeting with struggling students, calling parents, or going home before dark. Teachers who use AI aren't doing less work — they're redirecting work toward higher-impact activities.

The Tools That Help

• PPyGPT works well for teachers who want a desktop AI assistant that they can customize extensively for educational use cases.
• FFilesystem MCP lets AI interact directly with your local files — essay documents, rubrics, grade spreadsheets — without copying and pasting.
• FFlowise can build automated grading pipelines for standardized assignments like lab reports or math problem sets.

Starting Small

If you're a teacher curious about AI, don't try to revolutionize everything at once. Pick one pain point:

• If grading is killing you: Start with one assignment type. Let AI generate feedback drafts and see how much time you save.
• If planning is killing you: Start with one prep. Let AI generate a week of lessons and customize from there.
• If differentiation is killing you: Start with one activity. Let AI create leveled versions and see how students respond.

Each small win builds confidence and skill. By the end of a semester, you'll have an AI workflow that saves hours per week and improves the quality of what you deliver to students.

You became a teacher to teach. AI can help you get back to that.