What teachers need before students meet AI
Every national rollout that worked trained teachers before students got the tools. The ones that collapsed did it the other way round.
The closest evidence we have is about rolling out AI tools, not about teaching students what AI is. It points the same way. I want to be straight that it is not the same question, and that the research on teaching AI itself is thin: one systematic review found 22 empirical studies and no longitudinal ones at all.
South Korea spent roughly $460 million on AI-powered digital textbooks and pulled the programme after about four months of backlash over inaccuracies, data privacy, and teacher workload. The Ministry of Education's own report, published February 2025, found no significant improvement in academic performance. The diagnosis was consistent across the coverage: the technology shipped before the teachers were trained.
Estonia's AI Leap ran the opposite sequence. The Ministry of Education and Research piloted it from 1 September 2025 with teacher training first and student tools second - 3,000 teachers, 20,000 students.
Meanwhile only 18% of US teachers have any formal guidance on AI at all, and 34% have none (RAND / Gallup / Walton, "Teaching for Tomorrow: Closing the Expectations Gap," fielded 9 February to 2 March 2026, n=2,069). Teachers in higher-needs schools were less likely to have guidance than those in wealthier ones.
I'd like to talk to teachers
I've been reading everything I can find on how AI gets introduced in schools, and one pattern keeps repeating. The national rollouts that worked trained teachers first and let students near the tools second. The ones that collapsed did it the other way round - South Korea spent about $460 million learning that, and pulled the programme in four months. Meanwhile only 18% of teachers here have any formal guidance on AI at all, and a third have none.
What I can't get from reading is what any of that looks like in a real classroom this year. So I'm asking instead. Thirty minutes, no pitch, no pricing, nothing to buy at the end - I'm not selling anything on these calls and I'll say so again when we start.
Roughly what I'd ask about:
- What your school has actually told you about AI, if anything, and whether it was useful.
- What you had to work out on your own - and what you still haven't.
- What your students are already doing with it, and how you found out.
- What you'd want to be true before AI gets introduced to them deliberately.
In return: everything I learn goes back out as free write-ups and guides on this site, and you get the draft before anyone else does. If you'd rather skip the call and just read the results, that's fine too - they'll be free either way, with no email required.
Set up a time to talk30 minutes · not a sales call · nothing to buy
Findings
Nothing published yet. The conversations start first, and what comes out of them gets written up here. If you want it when it lands, the list below is the only way I will tell you.