by Rebecca Thomas
Just Because We Could, Does It Mean We Should?
No doubt you’ve been in a room full of teachers when they hear that AI can save time, increase efficiency, and even mark writing with a rubric for you. Smiles, grins, excitement, and "show me how" follow quickly after. I too am guilty of that reaction.
Then comes the warning about privacy and data settings.
Then the debate about which app is best; Papyrus seems to be a clear favourite for some.
Who can blame educators? At least if the AI does it, we get consistency across the school. We know the drill—some schools have those ‘harsh’ markers, others those lenient ones, and everything in between.
Some agree that using AI as a critical friend, or to give feedback, is a good use of time.
But here’s the part that makes me pause.
AI can recognise patterns, match data, and align student writing to a rubric. But writing is more than just structure and patterns — it’s expression, intent, and nuance.
AI doesn’t know when a missing full stop is an intentional choice rather than an oversight.
AI can’t sense when sentence structure is crafted for effect rather than a mistake.
AI doesn’t weigh the subtle difference between “most” and “many” in a rubric quite like a teacher does.
I’ve just spent two robust hours moderating writing across six schools (one high school, five primary). All of us are seasoned, long-in-the-tooth moderators. All of us are eager to examine the capability of our writers.
The first moderated piece is usually the longest and most thorough. The lengths we go to justify and discuss the impact of one mark is well-orchestrated and passionate, but at the same time, during the battle of wills and justification, eventually something resonates with us, and we see another’s point of view. When we hear the alternative perspectives, we can be talked around, and we use the battle of wills to reflect on our ability to analyse even the tiniest of details and share a window that the initial marker may have missed on first glance.
It’s tiring, but always something to be learned.
Eventually, the process gets slicker and faster. We dust off the cobwebs, become more motivated, and more confident. We usually return to the initial discussion we procrastinated over when we get stuck.
Despite AI’s ability to do this at speed and consult hundreds of samples it now probably has logged into its language model, I doubt it questions itself so much. I doubt it reflects and grows with the process like we do as teachers.
During one of our conversations we started discussing whether our students are using AI more, and if we notice it in their writing. Despite the reassurances some schools gave about having honesty policies or being locked down with platforms like Hāpara, we’d all be fools to believe it’s true.
If AI can deceive humans and generate images of Donald Trump DJ-ing with Putin, don’t think AI isn’t being used by our students to hack our school systems. Yes, of course they cheat and beat the system (even my son has revealed how) —because nothing says creativity like a 15-year-old using a VPN to bypass school restrictions and asking an AI chatbot to fix the issues that come with it. And just as easily as our schools try to lock things down, the kids outsmart us again; cat and mouse begins.
Our students are not just finding creative ways to break the system; they’re turning it into an art form. If there’s a hack, there’s a student ready to use it—AI is their sidekick, helping them make homework look like their own, while they’re off hanging out with friends. But sure, the monitoring systems are in place. All we need is to stare at the screen... don’t hold your breath, because the real activity has already been disguised.
And really, who could blame them? What would you have done if you’d been given a critical friend in the form of a bot to do your homework so you could hang out with your mates?
Who needs that extra stress when you can just ask the AI for help? Teachers use it for reports, emails and marking after all.
Evan a friend of mine recently pointed out when she was asked to do a character reference for another friend she just used AI to craft it.
At present, there are clues and breadcrumb trails for savvy teachers marking writing samples looking for evidence of AI use, but soon, if we keep feeding the algorithms genuine work samples from students, AI will get better at adding incidental mistakes to make the sample look human and authentic.
The Ministry of Education overviews of assessment collection and data gathering make me smile. In 2026, 80% of what we see online will most likely l be influenced by AI. So, is there really a place for e-asTTle written samples and STAR tests in 2026? Is that truly a good measure of what our kids can do? Or is it even a measure of what we need to prepare them for in the future?
The truth is, these assessments were planned in a time before AI was on the radar, before it was part of our everyday lives. Yet, these same assessments will soon be used to measure teachers’ success and school performance. As AI becomes more ingrained in our lives, we must ask ourselves: are these assessment schedules now outdated? What are we really measuring when we focus solely on traditional assessments? Are we preparing our kids for a world where "at" or "above" national standards doesn’t mean what it used to?
In a world where AI is ubiquitous, if students can use AI to do their writing, solve problems, hack our school security systems, or even complete assignments, how does that change what we should be measuring?
Are we preparing them for a future where AI isn’t just a tool, but a way of life?
In a sense, AI forces us to rethink the way we assess students. It challenges the value of the rubrics we’ve relied on for so long. If AI can nail the structure and grammar but miss the deeper human nuance, is that what we want to measure? It’s a big question.
AI is not just a tool for students, but for the entire educational system itself, maybe it’s time to rethink the very idea of "success"—because in the world of AI, does national "at" or "above" even exist anymore?
Comentários