What teachers actually want from AI (and what they don't)

Ask a room of teachers how they feel about AI and you'll get the same two answers at once. Relief, because some of this job is soul-crushing and a machine could have it. And dread, because every headline for two years has hinted that the machine is coming for the whole job, not just the bad parts.

Both reactions are correct. The problem is that the people building the tools keep aiming at the wrong half.

What do teachers actually want AI to do?

Teachers want AI to take the work that was never really teaching. Make the second version of the worksheet. Grade the multiple choice. Draft the parent email so they only have to fix it. Turn the PDF into something students can do online. Pull the numbers together so they don't spend Sunday copying scores between two systems. The boring, repeated, time consuming stuff that stands between a teacher and the actual kids.

What they don't want is for AI to do the teaching. They don't want it deciding what a student needs, or writing the feedback that's supposed to come from a person who knows the student, or standing at the front of the room. That part isn't busywork. That part is the job, and it's the part teachers are good at and actually like.

So the line is simpler than the headlines make it sound. Good AI for teachers does the work around the teaching. Bad AI tries to do the teaching itself. One hands you back your time. The other treats you like the part it's replacing, when you're the one part it can't.

That part matters, because a lot of the fear is aimed at something that isn't real. AI can't know that the quiet kid in the third row finally raised his hand today, or that this class needs the lesson slowed down, or why a parent is upset. It can generate a worksheet. It cannot read a room. The teaching, the human read of twenty-eight specific kids, is exactly the thing it's worst at, and exactly the thing it keeps getting marketed as able to do.

The honest version of an AI teaching tool is unglamorous on purpose. It doesn't promise to revolutionize your classroom. It promises to make the easier-and-harder versions so you don't retype them. To grade the parts that have right answers so you can spend your grading energy on the parts that don't. To handle the copying and tracking and converting that eat your evenings. It stays a tool. You stay the teacher.

That distinction sounds obvious, but most tools blow right past it, because "we'll do your busywork" is a smaller, less exciting pitch than "we'll transform education." The smaller pitch is the one teachers actually want. Nobody got into this job hoping a robot would take over the teaching. Plenty of people would take a robot that takes over the data entry.

That's the side we're building Worksheet Builder for. The making, the grading, the converting, the tracking, the work that was never the point, so the part that is the point stays yours. It's an assistant, not a replacement, and we mean that as a design decision, not a slogan. It's not live yet. The beta's coming, and the list goes first.

Want in early? →

Free during the beta.