Why your teaching instinct is your AI superpower

The teachers getting the most out of AI aren’t the keenest on it. They’re the most sceptical.

AI is already in classrooms across Australia and beyond. It’s not hard to see why: according to Atomi’s Tech in Schools 2025 survey, 60% of educators believe that AI edtech tools support their understanding of student progress, and the majority believe they decrease planning time.

But efficiency and time savings are just one part of the picture.

AI works in the classroom when teachers stay in control. They use their teaching instincts to be precise about how, when and where they let AI in.

In conversation with Sarah-Eleni Zaferis, this blog explores how AI turns feedback from a verdict to a real-time loop, what separates good from great AI usage in the classroom, and why your pedagogical experience and teaching judgement are your biggest assets in the AI era.

Watch Sarah’s full interview on YouTube.

Why the best teachers are the most sceptical about AI

Enthusiasm isn’t what makes a teacher good with AI. In fact, Sarah suggests the opposite: the teachers who get the most from AI are the most discerning.

“In my experience, the teachers that use AI the best are the ones that are most sceptical about it. They’re very precise about what they will let it near and what they won’t, and it’s that discipline that gets them the best results.”
— Sarah-Eleni Zaferis

This scepticism allows teachers to be clear about where AI fits into their workload. Typically, that involves streamlining the more tedious parts of the job: drafting lesson outlines, summarising curriculum changes or adapting materials for different learners.

AI isn’t a teaching shortcut. It frees up time for the parts of the job only a person can do:

• Noticing when a student has gone quiet, and may need more careful, one-on-one explanation of a new concept
• Reading the room in real time to know when to push a student further, or when to take a step back and offer more support
• Building trust with learners and creating a classroom environment where students feel safe to get something wrong, make a mistake and try again

“The best teachers are not just using AI to do their teaching, but rather they’re using it to free up the administrative load, their calendars, and giving themselves time back, and they’re using that time to do the most human elements of their job.”
— Sarah-Eleni Zaferis

From feedback as a verdict to feedback as a loop

Before AI, quality feedback was often rationed. Not because teachers didn’t care, but because the hours simply didn’t exist to provide specific feedback in real time.

A tick, a grade and a short acknowledgement were all most educators could manage. The real comments were saved for bigger assessment milestones, often arriving two weeks after submission, by which point the student had already moved on to a new topic.

What AI changes is timing. Feedback lands in the moment when it matters, with enough time for learners to revise and focus their studies ahead of exam season.

“Before AI, giving every student detailed feedback on every piece of work was impossible. What AI changes is frequency and immediacy. The shift is from feedback as a verdict to feedback as a loop: here’s what to fix, and an opportunity to have another go.”
— Sarah-Eleni Zaferis

Plus, AI personalises feedback in more detail than teachers could otherwise manage. It gives learners more than a mark. It shows them how to improve, and points them to the right revision.

But guardrails matter when AI delivers instant feedback to students. Off-the-shelf AI tools typically aren’t built with pedagogical principles or the oversight of academic experts. Instead, it’s worth finding AI tools designed for today’s learners.

For example, Atomi’s AI is trained, tested and evaluated based on thousands of Atomi questions and responses. Our in-house machine learning experts, academic content creators and high school teachers supervise it closely. This process enables Atomi’s AI to reliably deliver feedback that accelerates learning and meets users’ expectations.

Where else is AI making a real difference for teachers?

But marking and feedback are just one place AI helps learners. Sarah points to a few less obvious places where AI is changing day-to-day teaching for the better:

• AI lets teachers build multiple lesson plans from diverse learning materials (from videos to curated podcast playlists) or tailor exercises and revision resources to suit different needs, learning profiles and abilities.
• When a student doesn’t get it the first time, AI offers another analogy or angle on the spot to make new concepts stick.
• Teaching outside your specialist area? Common in secondary and regional schools, AI helps you get up to speed and sanity-check your understanding.

What separates good from great AI usage in the classroom

For Sarah, the difference comes down to answering this question: “What are you asking the tool to do?”

Good AI usage starts with seeing the tool as a way to save time. Great AI usage asks how the tool helps you teach better.

Again, it comes down to judgement. Great teachers have a clear picture of what good looks like, so they can direct the tool, review what it gives back and leave out what doesn’t hit the mark.

That judgement shows up most clearly with the students who are easiest to miss.

Most conversations about AI and differentiation focus on the students flying ahead or really struggling. These learners on the “edges” get the most attention because they have to. But Sarah flags another group of learners who stand to benefit from AI: students who sit in the “quiet middle”.

“There are kids who do exactly what’s asked: hand homework in on time, sit quietly, and never give you a reason to worry. And because they never get flagged, they end up getting the least of you. I always felt that deeply. What AI has been able to give those students is instant feedback, tailored practice questions, and extra support to help each student on their own learning journey.”
— Sarah-Eleni Zaferis

Get started with AI in your classroom

AI doesn’t replace a great teacher’s instinct and experience. Instead, it extends it. The teachers getting the most from AI are the ones who stay precise about what they hand over and what they don’t, and who use the time it gives back for the parts of teaching only they can do.

That’s the same standard Atomi holds its own AI to: built to support a teacher’s judgement, and never to bypass it.

If you’re thinking about where AI could support your own teaching, explore Atomi’s approach to AI in education to see what teacher-led AI looks like in practice.