AI Implementation in Schools: Where to Start
Most schools aren't ignoring AI. They just aren't sure where to start.
I've been having a lot of conversations with schools about AI, and one thing I keep coming back to is the need to separate two conversations.
How should staff use AI?
And:
How should students use AI?
They're related, but they're not the same conversation.
This article is about getting staff started.
For many schools, there are already teachers using AI tools to help with planning, administration, differentiation and other tasks. But if there aren't clear expectations around what staff can and can't do with AI, that experimentation often happens quietly.
Teachers might be using AI and learning useful things, but they don't necessarily feel comfortable sharing that with their colleagues.
That means schools can miss out on one of the biggest benefits of AI adoption: staff learning from each other.
Start with practical AI use cases
The first conversation I'd have with a school leadership team isn't "Which AI tool should we buy?"
I'd start with:
Where could AI actually be useful in our school?
Look at where staff are currently spending time.
Is it:
lesson planning?
administration?
reporting?
differentiation?
creating resources?
drafting communications?
organising information?
You don't need to find an AI solution for everything.
In fact, I think trying to tackle everything at once is one of the easiest ways to make AI implementation feel overwhelming.
Pick a few genuine use cases and start there.
If a teacher is spending hours on a repetitive task, that's a much better place to experiment than introducing AI simply because everyone is talking about it.
The question should be:
What problem are we trying to solve?
Then we can work out whether AI is actually useful for that problem.
Give staff permission to experiment
The second step is giving teachers time and permission to try things.
One-off AI professional development isn't enough.
A staff presentation can introduce possibilities, but people need time to actually use the tools.
They need to try something.
See what works.
Find out what doesn't.
Share their experience.
Try again.
This is particularly important because teachers will have very different levels of confidence with technology.
Some staff will already be experimenting with AI.
Others may not have used it at all.
A useful AI implementation strategy needs to give both groups somewhere to start.
And importantly, staff need to feel psychologically safe while they are learning.
If teachers aren't sure whether they are "allowed" to use AI, they may keep their experiments to themselves.
That means the school doesn't benefit from the knowledge being developed by those teachers.
Create an AI policy and clear boundaries
This is where school leadership needs to provide clarity.
Many schools don't yet have clear expectations around acceptable AI use for staff.
That creates uncertainty.
Can teachers use AI to help write a lesson plan?
Can they use it to draft a parent communication?
Can student information be entered into an AI tool?
Can AI be used when preparing reports?
What information should never be entered into an AI system?
Without clear boundaries, every teacher is left to work these things out individually.
A clear school AI policy can provide the framework staff need to experiment more confidently.
One approach I like is a traffic light model.
Green: What is okay?
These are uses that the school is comfortable with and may actively encourage.
Amber: What is conditional?
These uses may be appropriate in some circumstances but require additional care, checking or approval.
Red: What is off-limits?
These are uses that aren't permitted because of issues such as privacy, safety, accuracy or other school requirements.
The exact categories will be different for every school.
The important thing is that staff don't have to guess.
Policy shouldn't just tell staff what they can't do
There's another benefit to going through this process.
Writing an AI policy forces a school to think about the opportunities as well as the risks.
Instead of simply asking:
"What AI use should we ban?"
Leadership teams can ask:
"Where is AI genuinely useful, and where do we want to encourage its use?"
That might uncover practical opportunities that hadn't previously been considered.
It also gives staff a clearer sense of what the school is trying to achieve.
AI policy shouldn't become a long list of restrictions that nobody understands.
It should help staff make better decisions.
Focus on confidence, not compliance
One of the biggest mistakes I see in technology adoption is focusing too heavily on compliance.
Tell staff the rules.
Give them a policy.
Run a PD.
Tick the box.
But if teachers don't feel confident using AI, they aren't suddenly going to become confident because they've read a policy.
They need opportunities to practise.
They need to see what other teachers are doing.
They need permission to experiment.
And they need somewhere to ask questions when something doesn't work.
This matters for students too.
Teachers can't effectively help students navigate AI if they don't understand the technology themselves.
Staff AI use and student AI education are separate conversations, but they are connected.
If we want students to develop responsible AI habits, teachers need enough confidence and understanding to guide those conversations.
What should schools do first?
If your school is only just starting to think seriously about AI implementation, I'd suggest beginning with these four areas:
1. Identify useful AI use cases
Find the areas where staff are losing time or doing repetitive work. Start small and focus on genuine problems.
2. Give staff time to experiment
Don't rely on a single PD session. Give teachers opportunities to try AI, compare experiences and learn from each other.
3. Create clear boundaries
Develop a practical AI policy that explains what staff can do, what requires additional care and what isn't allowed.
4. Build confidence
Make AI professional development practical. Staff need to understand not just what AI can do, but its limitations, risks and where human judgement is still essential.
AI implementation needs to be practical
Schools don't need to have every answer before they start.
AI will continue to change, which means a school policy or strategy should be something that can evolve as staff learn more and new tools emerge.
The goal isn't to become an "AI school" overnight.
It's to start identifying where AI can genuinely help, give staff a safe environment to experiment and create enough structure that people know where the boundaries are.
And then keep learning.
That is a much more useful starting point than simply telling staff that AI is coming and hoping they work out what to do with it.
Helping schools get started with AI
I work with schools to build practical understanding around AI, including how staff can approach AI confidently and responsibly.
For schools developing their AI strategy, AI professional development for teachers can be a useful starting point for building shared understanding, identifying practical use cases and opening up conversations about appropriate boundaries.
The next step is making sure that understanding flows through to students too, so schools can develop a consistent approach to AI literacy across the community.
