For school leaders
Whole-staff AI professional development: what to do after the prompting workshop
Many schools have done their first round of AI training. Usually it's framed as foundation AI skills: a tools demo, a session on writing better prompts, and then a policy that sets out what's allowed. Staff leave curious and some start using AI for their planning. A term later, the classrooms look much the same. Leaders I talk to can feel that something is missing, and they're not sure what comes next.
The problem underneath
Prompting is a useful skill, and a policy is a necessary document. Neither of them changes the task, and the task is what decides whether students do their own thinking. If a chatbot can finish the task in one go, plenty of students will let it, however well their teacher can prompt. So the next step for staff is learning to design tasks where the thinking has to be the student's, and to recognise that thinking when they see it.
What the evidence says about PD that changes practice
The research on professional development is fairly consistent on one point. Practice changes when teachers try something in their own classrooms and get feedback on it over time. A large review of coaching studies found positive effects on teaching and on student results, although the effects shrank as programmes grew (Kraft, Blazar & Hogan, 2018). Online courses can work too, but the studies that show it involved sequenced, subject-specific courses over three semesters, about 100 hours in all (Boston College e-Learning for Educators). One session can't do that on its own.
Why a shared model outlasts any tool
AI tools change every term, so training built around one tool dates fast. A shared way of talking about students' thinking keeps working whatever the tools do. When the History department and the Biology department both ask for the same kind of evidence, students hear one message instead of several, and staff can moderate each other's tasks.
The Thinking Pendulum gives staff that shared language. It describes two ways students use AI: as a vending machine that hands over answers, or as a thinking partner that pushes their own thinking through three zones (sense checking, outcome determining and problem shaping).
What whole-staff PD looks like with the Thinking Pendulum
- A launch session introduces the model to all staff, using tasks from your own school.
- Each teacher redesigns one task they're already teaching this term, stage by stage, with the evidence each stage should leave.
- Departments moderate the evidence together: students' logs, their chat bibliographies and the process rows of the rubric.
- Repeat next term with another task, and move further along the pendulum.
It runs on the units your staff are already planning, so there's no new scheme of work.
How you'll know it's working
Look at what students hand in. If the keep or reject logs show real decisions, the chat bibliographies show the thinking moving between zones, and the process rows of the rubric are rising, practice has changed.
Questions school leaders ask
How much staff time does this take?
One launch session, then time teachers would spend planning anyway. Each teacher reworks a task they're already teaching, so the main cost is department time for moderation.
What if some staff are sceptical about AI?
That's fine. The Pendulum describes students' thinking, and the first zone, sense checking, works without AI at all.
Do teachers need to be confident AI users first?
No. The work is in task design and in reading the evidence, which are things teachers already do.