The practical AI update · Part 1

What Teachers Are Actually Using AI For

The useful question is not whether teachers use AI. It is which parts of the job become lighter without making the teaching worse.

There seems to be a new AI tool for teachers every week. Most of them arrive promising to transform education. I would settle for one that helps me finish Thursday’s work before Thursday night.

The strongest use cases are not mysterious. Teachers are using AI to prepare lessons, make or revise activities, differentiate materials, draft routine communication, build assessments, and turn rough notes into usable documents. A national Gallup survey found that weekly users estimated saving an average of 5.9 hours a week. That is not a guarantee. It is evidence that a repeatable workflow can give real time back.

The work I keep handing off

I have used AI to revise a syllabus, build a first unit, create emergency sub plans, draft parent communication, organize product files, turn a classroom idea into a working game, create vocabulary and morphology practice, and build a feedback system for essays. The biggest savings did not come from asking, “Make me a lesson.” They came from giving the tool the materials, constraints, and decisions I already had.

• Planning: organize standards, texts, time, and an end goal into a workable sequence.
• Adaptation: create another reading level, scaffold, example, or practice set from a teacher-approved source.
• Production: turn an existing plan into directions, slides, checklists, rubrics, or a printable handout.
• Feedback support: sort patterns, draft comments from a rubric, or identify where a response needs closer teacher attention.
• Communication: shape the context around the one sentence that needs to come from the teacher.

Teachers are also naming the failure points

Teacher conversations online are not one long endorsement. They are full of complaints about generic materials, factual errors, lifeless emails, shallow grading, and colleagues who use output they have not read. That criticism is useful. If a tool produces thirty questions but you spend forty minutes repairing them, it did not save time. It created a different pile.

The teachers reporting the most useful results tend to do three things: start with specific source material, ask for a clearly defined deliverable, and check the result because they know what quality looks like. The teacher is not removed from the process. The teacher moves from typing every word to directing and editing the work.

The line I will not cross

AI can help me see patterns in student writing. It cannot decide the final grade. It can draft a difficult family email from facts I provide. It cannot know the child or carry the relationship. It can generate examples. It cannot tell me whether an example will land with the students sitting in front of me at 1:40 on a Friday.

The rule is simple: hand off repetition, formatting, variation, and first drafts. Keep judgment, relationships, privacy decisions, and anything with consequences for a student.

Sources and further reading

• Gallup: Three in 10 Teachers Use AI Weekly, Saving Six Weeks a Year
• OpenAI: how educators are using ChatGPT for time-saving tasks
• Teacher discussion: where AI helps—and where it makes the work worse