TL;DR
YouGov asked 1,033 teachers about AI on behalf of BETT, for the third year running. Adoption is up sharply, workload is down for half of users, and hours worked have barely moved. Expectations have fallen faster than usage has risen.
The gap between workload and hours
Eight in ten now use AI for their work, and weekly use has doubled to four in ten. Among users, 51% say it eased their workload; 40% saw no change and 9% found it made things worse. But only 35% work fewer hours as a result, while 55% put in the same time as before.
Regular users typically recover one to two hours a week. The survey’s explanation for why that does not show up in hours is the one every sector eventually reaches: the time saved gets absorbed by other demands.
BETT’s Duncan Verry described enthusiasm as cooling. Last year 68% expected to increase their use over the following twelve months. This year that figure is 47% — a 21-point fall in a single year, against rising adoption. His summary was that teachers “have taken AI about as far as they can on their own”.
Support is where the divide sits
Two in five report no training or support at all from their school. Almost half say there is no written AI policy, and where one exists only 18% say they understand it well. Independent schools do better: 61% of their teachers had training, against 41% in the state sector.
Usage clusters around preparation rather than assessment — 76% for lesson resources, 39% each for parent correspondence and pupil reports, and just 8% for marking. That last figure is notable given yesterday’s research finding that AI marks essays generously and errs most on the weakest work.
Looking forward
The headline finding generalises well beyond schools. A tool that reliably saves two hours a week produces no measurable hours reduction unless something is deliberately removed from the workload — otherwise the gap simply refills. That is a management question, not a technology one, and it is the same paradox showing up in UK productivity statistics.
The 21-point collapse in expectations deserves attention too. This is what the plateau looks like when it arrives.