Everyone Agrees Marking Is Breaking Teachers. The Argument Is About the Wrong Fix.
There's something about marking that teachers' unions and school leaders genuinely agree on: the load has become unsustainable, it leaks into evenings and weekends, and it's wearing good people out. I build AI marking software, so I'm not a neutral party on the solution. But you don't need to take a position on the wider workload debate to see that this shared ground is being spent on the wrong question.
A note on bias. I run an AI marking company (Top Marks AI), so treat what follows with appropriate scepticism — I have an obvious interest in where this lands. What this piece is not about is who should win any particular dispute. It's about the problem almost everyone in education already agrees on, and the figures are the profession's own.
- On marking workload, there is broad agreement across the profession: teachers' unions and school leaders alike accept the load has become unsustainable.
- This isn't a soft claim. The average teacher spends around 230 hours a year marking (UCL) — the single biggest discretionary slice of the 1,265 directed hours, and a leading driver of teacher attrition.
- That agreement gets squandered on a bad binary: teachers absorb the hours, or the work doesn't get done. Both ends hurt — and in the second, it's students who lose feedback first.
- The way out follows from something everyone actually accepts: feedback is the value; hand-marking stacks of scripts at the weekend is the grind. They are not the same thing.
- Keep the feedback, remove the grind — with the teacher's judgement still on top. That's the rare move that's good for staff, good for leaders managing fixed budgets, and good for students at once.
Where the profession agrees
Ask a classroom teachers' union, a leaders' body, and a head trying to hold a timetable together inside a shrinking budget what's wearing teachers down, and on marking you'll hear a strikingly similar answer: there's too much of it, it leaks into evenings and weekends, and it's exhausting good people. That kind of shared ground is worth building on.
The figures bear it out. A UCL study found the average teacher spends around 230 hours a year on marking — the largest single block of discretionary time in the job, and the part that most reliably eats into Sundays and the back end of half-term. "Drowning in a sea of marking" isn't rhetoric; it's one of the most consistently cited reasons teachers leave, and replacing a teacher costs a school somewhere around £10,000–£15,000. So this is a wellbeing problem and a retention problem with a price tag attached — and on its scale, there's no real disagreement.
The wrong question
Here's where that agreement goes underused. The debate tends to collapse into a single question — who has to absorb the hours? — and framed that way, every answer is a bad one. Either teachers keep soaking up unpaid evenings and weekends, which is unsustainable and a slow leak on retention; or some of the marking simply doesn't happen.
And when marking doesn't happen, it isn't shared out evenly: it's the students who most need timely, specific feedback who tend to lose it first. "Whether a piece of work needs marking" can quietly drift toward "the bottom set gets less," which is the opposite of what anyone in this conversation actually wants. The reason that binary feels inescapable is that it treats marking as one indivisible lump of work — something you either do or don't do. It isn't.
The distinction that unlocks it. Even the union guidance that prompted this latest round of coverage makes the point: feedback "does not necessarily mean marking has to be in written form, or that it should be done in a prescribed manner, for example, using different coloured pens." That isn't a controversial position — it's close to common ground. Feedback is the value. A teacher hand-marking in green pen at 9pm on a Sunday is the grind. Those are two different things, and almost everyone agrees they are. The debate just keeps treating them as one.
The fix the agreement points to
Once you separate the value from the grind, a third path opens up that isn't "teachers mark everything at the weekend" and isn't "students stop getting feedback." It's: keep the feedback, remove the grind.
That's the entire reason AI marking exists. Give a teacher a fast, consistent first draft of the marks and feedback — aligned to the exam board's assessment objectives, not generic prose — and you hand back the hours without taking anything away from the child. Our own studies put that saving at roughly a 55% reduction: about 125 of those 230 hours returned per teacher, per year. Across a department, that's well over a thousand hours redirected to planning, intervention and actual teaching.
The non-negotiable is that the teacher stays in the loop. No AI mark of consequence without a teacher's eye on it. AI should replace the repetition in marking, not the expertise — and not the wider human role of being a teacher who notices when a student's writing is telling you something the mark scheme can't. Used that way, it isn't deskilling and it isn't automation for its own sake. It's the rare move that serves teachers, the leaders managing the budgets, and students, all at once — which is exactly what you'd want a fix to do, given how much agreement there already is about the problem.
Where that leaves schools
The wider workload questions will be settled where they're always settled, between unions and employers. But the most productive thing a school can do this term doesn't depend on any of that. It's to ask a simpler question, one that sits squarely inside the agreement: how much of our marking load is genuinely about helping students improve, and how much is just hours of hand-transcribing marks that a well-built, teacher-supervised system could give back?
Get that right and the weekend-marking problem starts to soften on its own — not because anyone won an argument, but because there was less grind left to fight over. If you want to see how the accuracy holds up against exam-board standards, our secondary-schools guide works through the evidence, and how to choose AI marking software sets out the questions to ask any vendor — including us.
See what handing the hours back looks like
If your staff are losing evenings and weekends to marking, we'll show you the accuracy data for your specific subjects and exam boards, and how a teacher-in-the-loop workflow gives the time back without giving up the feedback.
Frequently Asked Questions
How much time do teachers spend marking?
A UCL study found the average teacher spends around 230 hours a year on marking — the single largest discretionary slice of the 1,265 annual directed hours, and one of the most consistently cited drivers of teacher attrition. Replacing a teacher costs a school in the region of £10,000–£15,000, so the marking load is both a wellbeing and a retention issue.
If teachers do less marking, do students lose feedback?
That's the risk in framing it as "mark everything or don't." It's lose-lose: the students who most need timely feedback are the ones who stop getting it. The more constructive position is to separate the value (feedback that helps a student improve) from the grind (a teacher hand-marking stacks of scripts at the weekend). AI marking lets schools keep the first and remove the second, with teachers reviewing and overseeing the output rather than producing every mark by hand.
Does AI marking replace teachers?
No. AI should replace the repetition in marking, not the expertise. A teacher reading a script catches things a mark scheme can't — including safeguarding signals — and the accountability for any mark of consequence stays with them. Used well, AI marking takes the mechanical hours off a teacher's plate so they can spend more time on planning, intervention and the parts of the job that actually need a human.
Is AI-generated feedback as good as a teacher's?
The reliable model isn't AI instead of a teacher — it's AI as a fast, consistent first draft that a teacher reviews and refines. Good AI marking is calibrated to the exam board's own assessment objectives rather than producing generic prose, and it should surface its own uncertainty so teachers can focus their attention where it's most needed. The teacher's judgement is what turns a draft into feedback a student can act on.
We use cookies for analytics and marketing to improve your experience — these are only set if you accept. Decline and we'll only use cookies that are strictly necessary. (Live chat is always available either way.) Learn more in our Cookie Policy.