If We Stop Owning Our Words

We’re starting to talk about large language models as if they understand us.

They don’t.

They recognise patterns. They stitch together likely sentences. They echo our rhythms and styles. They produce what looks like meaning, but without intention or lived experience behind it.

That matters.

When syntax starts dictating semantics, reading and writing stop being acts of meaning-making and turn into quick transactions with an algorithm.

Language Is More Than Structure

Language carries identity, memory, and contradiction. It holds the cultural baggage we inherit and the stories we choose to tell. None of that is transferable to a machine.

LLMs can shuffle words, but, if we let them define our language, we risk hollowing out the part that makes us human.

Hand over language and you start handing over culture too. People in power have known this forever; change the language and you change the people.

But this shift isn’t slow and generational. It’s immediate. And we might not claim it back.

The Prompt Isn’t the Point

That said, I’m not anti-AI. I use it daily to test arguments, challenge my thinking, and try out alternatives. It’s useful - when you’re in charge.

But our questions need to be better. Not “What will the machine say?” but “What do I think, and how can this tool help me say it better?”

Students need to learn how to interrogate, challenge, and push back. We should be demonstrating how to refine questions and stress-test ideas - not to avoid thinking altogether.

This is why teachers aren’t becoming obsolete. Like all the times before, their role shifts. They become the keepers of nuance, voice, and collective cultural memory. The people who help students notice the difference between genuine meaning and machine-generated fluency.

We can help them make that shift. We can turn to in-class writing, more voice-led work, and more focus on process, to keep human voices at the sharp end of education.

But ultimately students have to want to think for themselves. That's a cultural shift in education, not just a curricular one.

The Line We Can’t Afford to Cross

If we continue to slip and let syntax replace semantics, we lose our ability to claim meaning.

We lose stories, perspectives, and the right to say, “This matters because I say it does.”

AI can support meaning-making, or it can flatten it. The difference lies in who’s steering. Who’s wrestling with ideas. Who makes the final call on what gets said and why.

The line is simple.

Students that don’t ask “What will the machine say?”...

…but “What will I say?”

That’s how our words stay human - even when machines walk and talk among us.

And if we don’t pay attention, the risk isn’t that AI will ‘take over.’ It’s that our own voices will thin out under the hum of the servers. [ITL]

Rory File is Head of Education and Partnerships for educational digital resources company ILT Education.

Rory File, Head of Education and Partnerships, ILT Education.