Why chasing AI misuse is the wrong fight — and what schools should focus on instead
While marking coursework recently, I found myself raising an eyebrow at a couple of short sections in a 100+ page document. The writing style shifted subtly — polished, oddly general, and strangely impersonal. It had that tell-tale “AI gloss” that makes you wonder whether a student had been lured into thinking ChatGPT could write their coursework for them, rather than with them.
And this raises a bigger concern.
If such work were ever challenged by an exam board, even if we couldn’t prove AI use conclusively, it could trigger a lengthy investigation — one that eats up time, energy, and goodwill. Worse still, it casts doubt on a student’s authenticity, whether or not any intentional wrongdoing occurred.
Detection tools: a poor solution to the wrong problem
It’s understandable that teachers feel uncertain — or even anxious — about the growing presence of AI in student work. We’ve spent years developing our professional judgment, and now we’re being asked to trust that judgment against tools that can produce polished work in seconds.
AI detection tools promise to help, offering probabilistic guesses based on linguistic patterns, stylistic anomalies, and statistical markers. But even when they work, the result is only a suggestion, not evidence.
And often, they don’t work — especially for:
- Students who are neurodiverse
- Those with English as an additional language
- Learners who write in highly structured or repetitive ways
These tools can unfairly penalise the very students we aim to support. Worse, they risk undermining trust and generating false positives that distract from real teaching and learning.
While some schools may still consider using detection tools as a temporary aid, they must be treated with caution and never mistaken for definitive proof. Their use is a short-term crutch — not a long-term solution.
Trying to catch AI misuse through detection is like using a speed camera to catch invisible cars. It might feel proactive, but it won’t get us where we need to go.
So what should we do instead?
We need to stop trying to catch AI use and start teaching it — ethically, intelligently, and with purpose.
That means:
- Empowering students to use AI tools for brainstorming, drafting, clarifying, and reflection — not deception
- Teaching proper acknowledgement of AI-generated assistance, just as we would with websites, books, or human help
- Adapting assessment to reward process over product: portfolios, drafts, planning documentation, and reflective commentary
- Integrating oracy: Expect students to explain, defend, or discuss their work to demonstrate understanding and ownership
Rather than policing AI use after the fact, let’s build a learning culture that values transparency, creativity, and critical thinking — one in which AI is a tool to enhance student thinking, not replace it.
A final thought (and yes, a little irony)
Yes — this post was drafted with the help of AI.
There’s something deliciously ironic about using AI to write a post about not using AI to catch students using AI. But maybe that’s exactly the point. AI, like the internet before it, isn’t going away. The real task isn’t to fear it or ban it — it’s to help students learn how to use it responsibly, reflectively, and skilfully.
If we want our students to navigate the future with integrity and adaptability, we must start modelling that ourselves — not by fighting AI, but by leading the way in using it thoughtfully.
