A teacher's workday with AI help

The OECD TALIS study shows that the average Estonian teacher works 46 hours per week. (Source: OECD TALIS, 2018) Of that, ~50% goes to things that aren't teaching: preparation, grading, bureaucracy.

This is a problem. Not because teachers shouldn't prepare. But because a large part of that time is repetitive work that a machine could do.

Here's one real day. Not an ideal - but what I've actually seen working.

Morning 7:30 - Preparation

Situation: You have 3 classes today. In 6th grade, three students need simpler material (gap from previous lesson). In 9th grade, you want to do a review test before the exam.

Traditionally: - Search for suitable material - Adapt it manually - Create test questions

Time: 45-60 minutes.

With AI: Prompt 1: "I have this worksheet on fractions. [Copy content] Make a simpler version with only same-denominator addition. 5 problems, increasing difficulty."

Prompt 2: "Generate 10 review questions for 9th grade chemistry topics: acids, bases, pH. Multiple choice, 4 options each. Include answers separately."

Time: 15-20 minutes (including review).

Savings: 30-40 minutes.

Day 10:00 - During class

Situation: One student asks something you can't explain well immediately. ("Why is water's pH 7, not 0?") Another finished quickly and is bored.

Traditionally: - Try to explain, comes out awkward - Tell the quick student "wait a bit"

With AI (if you have phone/computer): Prompt 1: "Explain why water's pH is 7, not 0. Explain so a 14-year-old would understand. Use a simple example."

Prompt 2: "Give 3 extra problems on pH calculation. More complex than usual."

Time: 30-60 seconds.

Note: This isn't for every lesson. But it's an option that didn't exist before.

After lessons 15:00 - Grading

Situation: 22 essays. Each needs feedback: what's good, what to improve, how to move forward.

Traditionally: Read essay, think about feedback, write. Repeat 22 times. Time: 10-15 min per essay = 3.5-5.5 hours.

With AI: Prompt: "Read this essay [paste text]. Give 3 feedback points: 1) what's good, 2) what needs improvement, 3) one question to make the student think further. Be specific and reference the text."

You read AI response, adjust (~30% of cases change something), add personal touch.

Time: 4-6 min per essay = 1.5-2 hours.

Savings: 2-3 hours.

Evening 19:00 - Week planning

Situation: Next week you want to do a project. Need a plan: what, how, how many lessons, how to assess.

Traditionally: Search for examples, think, create yourself. Time: 45-90 minutes.

With AI: Prompt 1: "8th grade, chemistry. Want to do a mini-project on acidity, 3 lessons. Suggest 3 different formats: one lab work, one research, one practical application."

Prompt 2: "Choose option 2. Create: 1) lesson breakdown, 2) required materials, 3) assessment rubric with 4 criteria."

Time: 20-30 minutes.

Savings: 30-60 minutes.

Day summary

That's one day. There are 5 days in a week.

What AI DOESN'T do

1. Doesn't decide the grade

AI can help with feedback, but final grade is your decision. You know context AI doesn't. AI in assessment covers this in depth.

2. Doesn't replace relationship

If a student needs a conversation, not feedback - that's your job. AI doesn't feel emotions.

3. Doesn't know everything

AI makes mistakes. It can say wrong things. You must verify.

Practical start

1. Choose ONE thing: preparation OR grading OR planning
2. Try for a week
3. See if it works
4. If yes - add next thing

Don't try everything at once. That's a recipe for burnout.

Summary

AI doesn't change teaching. But it changes how much time goes to things that aren't teaching.

3-5 hours per week is 150-250 hours per year. That's time that can go where it actually belongs: to students, to professional development, to rest.