Most teacher AI training is teaching the wrong thing.

There's good news and bad news in the way schools are training teachers on AI. The good news, according to Education Week's reporting on its own research centre surveys this spring, is that far more teachers are now receiving some form of AI professional development than were two years ago. The bad news is what that development tends to cover.

The pattern the reporting describes will be familiar to anyone who has sat through a staff-meeting demo: a session on generating lesson plans, drafting parent emails and speeding up admin. Useful, certainly. But it is training in efficiency, not in teaching. It leaves the actual craft — how AI might change the way a concept is introduced, differentiated, practised and assessed — untouched.

Efficiency is the entry point, not the destination

We start with efficiency too. It builds confidence quickly, and a teacher who has just saved an hour is a teacher who is willing to go further. The mistake is stopping there. In our AI professional development the efficiency module is one of four, and it is deliberately the second, not the last. The modules that follow — ethics and assessment, then personalisation — are where the classroom actually changes.

• Awareness: what the tools do, and hands-on experimentation without stakes.
• Efficiency: prompting, quiz generation, automating the repetitive work.
• Ethics and assessment: bias detection, academic integrity, and designing tasks that AI can't simply complete for a student.
• Personalisation: scaffolded materials, language support and accessibility — the part that helps the learner in the back row.

The other half of the problem is leadership

Education Week's piece quotes Michael Horn on what happens without a clear stance from the top: confusion. Teachers who don't know what the school's position is will default to caution, which usually means doing nothing. A short, plain policy — what's encouraged, what's off-limits, what students may use for which kinds of work — does more for adoption than another tools demo.

It's also worth being honest that plenty of teachers are not enthusiastic, and that scepticism is not a problem to be trained away. The most persuasive framing we've seen, echoed by a principal in the same report, is simply that this is about preparing students for what comes next. That is a goal teachers already hold. Good training connects to it rather than talking past it.

If you're choosing a programme, ask one question of the provider: what does a teacher walk away with that they can use on Monday? If the answer is a slide deck, keep looking. If it's prompt banks, assessment redesigns, classroom templates and a person to call when it doesn't work — that's training.