Something shifted in British education policy this month, and almost nobody noticed. The UK government is spending £23 million on AI “tutors” to reach 450,000 pupils — not because AI is good for learning, but because there aren’t enough teachers.
This is different. Every previous government AI-in-education initiative has framed technology as a supplement to human teaching: personalised learning tools, marking assistants, lesson planners. The UK is now the first major government to explicitly position AI as a substitute for missing staff.
The Numbers Behind the Crisis
Britain’s teacher shortage has been building for years, but 2026 is the year it became impossible to ignore:
- 23% drop in trained teacher recruitment this year alone
- Nearly half of secondary schools reported vacancies last year
- £1.4 billion spent on supply teachers last year — a band-aid on a haemorrhage
- Subject-specific gaps in maths, physics, sciences, and modern languages hitting deprived areas hardest
- Schools in up to 40% of secondaries are running deficits, forcing redundancies
The Department for Education has a target of recruiting 6,500 new teachers by 2029. They don’t have a coherent plan to hit it. Unfunded pay rises (2.8% in 2025), National Insurance hikes, and crumbling school buildings make the job progressively less attractive.
So instead of fixing the supply problem, the government is filling the gap with software.
AI Tutors: Supplement or Substitute?
Education technology company Efekta, which counts former Deputy Prime Minister Nick Clegg on its advisory board, is already deploying AI-led instruction platforms at scale internationally. Its system combines AI-led lessons with teacher oversight — students work at their own pace while teachers monitor progress.
“The system has always been designed with the human teacher at heart, and the teacher remains in control,” said CEO Stephen Hodges.
But when pressed on what happens when there isn’t a teacher — as in São Paulo, where 95% of English teachers don’t speak English — Hodges described a role shift. “If the teacher doesn’t speak any English, they become a motivator and a coach.”
That’s a hell of a reframe. A motivator is not a teacher. A coach is not an educator. What we’re watching is the semantic downgrading of the teaching profession to make AI deployment sound palatable.
Efekta claims 4.5 million students worldwide use its platform, with test score improvements of 25-30% in some deployments. Those numbers deserve scrutiny — “some deployments” is doing a lot of work in that sentence.
The Policy Vacuum
Here’s what makes this reckless: 76% of UK teachers now use AI daily, up from 53% last year. But 49% of schools have no AI policy — for staff or students. Two-thirds of secondary teachers report declining critical thinking in pupils who rely on AI.
The government is simultaneously banning smartphones in schools while spending millions on AI tablets to replace missing teachers. If that sounds contradictory, it’s because it is.
As we covered in our UK teacher AI confidence gap analysis, 60% of UK teachers use AI but 43% rate their confidence at 3 out of 10. You’re handing 450,000 pupils to technology that the adults in the room barely understand — and the adults in the room are increasingly not even there.
The New Zealand Lens
New Zealand is watching this unfold with particular interest. NZ faces its own teacher supply crisis — particularly in rural schools and STEM subjects. The Education Review Office has noted growing interest in AI-assisted learning, but policy is still in the “exploration” phase.
The UK precedent matters because it creates a template. If Britain can point to AI tutors as a “solution” to teacher shortages, pressure will mount on other countries with similar crises to follow suit. The question isn’t whether AI can deliver lessons — it’s whether we want a system designed around technology availability rather than educational evidence.
New Zealand’s smaller scale and stronger community-based school culture could make it more vulnerable to quick-fix solutions. Or it could make it easier to resist them. The choice matters.
What This Is Really About
The UK isn’t investing in AI tutors because the technology has proven itself. It’s investing because the alternative — paying teachers properly, fixing school buildings, and making the profession attractive — is more expensive and less politically convenient.
Education secretary after education secretary has tried and failed to solve the recruitment crisis. AI offers the appearance of action without the political cost of actual investment. It’s the same logic that gave us Meta’s AI code quotas replacing developers and Oracle’s 30,000 layoffs funding data centres — technology as a cheaper substitute for humans, justified after the fact by efficiency rhetoric.
But classrooms aren’t server racks. You can’t replace a physics teacher with a chatbot and call it education. You can call it content delivery. You can call it personalised learning. You can call it whatever Nick Clegg wants you to call it. But if the teacher isn’t in the room, the teacher isn’t teaching.
🔍 THE BOTTOM LINE
The UK is running a live experiment: can AI tutors fill the gap left by a broken teaching pipeline? The £23M bet says the government thinks they can — or at least that it’s cheaper than trying to recruit 6,500 real teachers.
The rest of the world, including New Zealand, should watch the results carefully. Not the test scores. The classrooms. Because the data that matters won’t be in the assessment reports. It’ll be in the rooms where a screen replaced a person, and nobody asked the students if that’s what they wanted.
Sources
- City AM — UK eyes AI in classrooms as schools face staffing strain
- Department for Education — Teacher recruitment targets and vacancy data
- Tes Magazine — Teachers using AI in daily work survey
- X/Twitter — Multiple threads on UK teacher shortage and AI policy
