Ask a teacher where their evenings go and a lot of them will say the same word: marking. A stack of assignments, the same rubric applied for the thirtieth time, the same handful of comments written out again in slightly different words. It's work that has to be done, has to be fair, and has to be consistent — and it arrives in volume, at the end of a day that was already full of teaching.
It's also, for exactly those reasons, a good candidate for a careful first pass. The emphasis is on both words: careful, and first.
What this is — and what it absolutely isn't
Let's be clear up front, because this is the part that matters most. AI-assisted marking does not mean a machine decides students' grades.
It means a consistent first pass against the teacher's own rubric, with drafted feedback, handed to the teacher to review. The teacher reads it, adjusts it, and owns every mark. The work just starts further along than a blank stack of papers. A version that quietly assigned final grades would be the wrong thing to build, and we wouldn't build it.
So everything here assumes the teacher stays in charge of every result.
Why a first pass helps even though the teacher re-checks it
If the teacher reviews everything anyway, what's the point? The point is that reviewing is faster and more consistent than starting cold:
- Consistency. Applying a rubric identically to the first paper and the fortieth is exactly where human attention frays. A first pass applies the same standard evenly, and the teacher corrects from there.
- The blank-page tax. Writing feedback from nothing is slower than editing a draft that's already in roughly the right place. The teacher shapes it into their own voice rather than composing each comment from scratch.
- Attention where it counts. The borderline cases, the surprising answer, the student who needs a real conversation — those get the teacher's full attention, because the routine 80% isn't draining it.
The result isn't less teacher involvement. It's the same involvement, spent on judgment instead of repetition.
Built on the teacher's own rubric
This only works if the first pass reflects this teacher's standards. Built on the teacher's own rubric and standards, the draft starts in the right place and sounds like the right subject. Built on a generic grading model, it would impose someone else's idea of what good work looks like — which is precisely what a teacher can't accept. The tool's job is to apply the teacher's standard consistently, not to invent one.
The teacher owns the final mark
It bears repeating, because it's the whole ethic of the thing: the teacher owns the final mark. The benefit is more time for teaching and less for the stack — not less control over what a student is told about their work. A student is still graded by their teacher; the teacher just didn't have to start from a blank page at 9pm.
Where Honewright fits
AI-assisted marking is one of the things we're developing: a consistent first pass against the teacher's own rubric, with drafted feedback for review. The teacher reads, adjusts, and owns every mark — the work just starts further along. Built on the teacher's own standards, and kept in Canada.
If marking is eating your teachers' evenings, tell us what's eating your time and we'll tell you honestly what it would take to give some of it back.