A recent news story from The Guardian exposed an academic for allegedly using AI to write an SMH opinion piece warning students against using AI.
Predictably, people had a field day. Including me.
In fact, I was so delighted by the irony that I rushed to share the story on LinkedIn… without reading the whole article. (You know the drill. Great headline. Strong opinion. Share button. Read later.)
Don’t judge me, okay? The timing was just too good.
You see, the story broke on the very same day as our long-awaited webinar, The Dark Side of AI Writing. So it felt like the universe had handed me the perfect case study: Academic warns students not to use AI. Academic gets caught using AI. What more proof did I need?
Quite a lot, it turns out.
Because when I went back the next day to read beyond the headline, I felt suddenly sheepish – since I’d joined the pile-on without having all the facts.
For starters, I never actually read the academic’s original opinion piece. It had already been taken down by then. So despite confidently weighing in on the quality of her AI-assisted writing, I had never seen the writing itself.
Not exactly a solid basis for a strong opinion.
After all, if I’m going to criticise someone’s use of AI, surely I should look at the output. Was it insightful or shallow? Original or generic? Well argued or full of holes?
I couldn’t possibly answer any of those questions without seeing the article. Yet there I was, happily passing judgement.
Which is when it dawned on me that I’d been focused on entirely the wrong thing.
We’re all so busy arguing about whether people are or aren’t using AI to help them write. Meanwhile, millions of people are already using it every day – and they’re going to be using it even more tomorrow. The genie’s out, the horse’s bolted, the ship’s sailed. Pick your cliché… because frankly, I think they’re all at the pub together by now.
So can we stop treating ‘Did they use AI?’ as the most important question in the room? I’m much more interested in how it was used. And ultimately, whether the final piece stood up on its own merits.
But that’s not the only thing that gave me pause about this particular case.
The academic was reportedly ‘caught’ by an AI detection tool that flagged the opinion piece as AI-generated. But these tools don’t exactly have a flawless reputation. Some have famously flagged works by Jane Austen as AI-generated. So if a nineteenth-century novelist can trigger an AI detector, I’m not sure we should treat the results as irrefutable evidence.
But then again, the detector only tells us that AI may have been involved. It doesn’t tell us how it was used.
According to reports, the academic uploaded around 40,000 words of her own research and ideas into Microsoft Copilot, which then helped structure and draft the article. The university defended that as an appropriate use of the technology – and on reflection, I think they have a point.
Feeding your own research and thinking into a tool to help with structure and clarity is closer to hiring an editor than buying a ghostwriter. The ideas were hers. The intellectual work was hers. AI handled some of the scaffolding. But that’s a very different scenario from someone typing a prompt into AI, copying the output and passing it off as their own work.
And let’s not forget that we’ve been using tools to improve our writing for decades. Spellcheck. Grammarly. Google. Thesauruses (remember those?). Sure, there’s a big difference between a tool that corrects your words and one that helps shape them. I’m not pretending otherwise.
But the question was never really about the tool. It was always about the thinking behind it. Did the writer share genuine ideas and turn them into a coherent argument? Or did she hand all of that over?
That’s the line that matters.
Consider how we think about other workplace tools. Nobody looks at a polished bar chart and says ‘it would have been more credible without Excel’. Nor does anyone get extra points for designing their social media graphics without Canva.
Writing deserves the same treatment. Yet the moment the letters ‘A’ and ‘I’ appear together, we react as though the writer has outsourced their soul.
To be clear, I’m not giving anyone a free pass to hand their work over to ChatGPT and call it a day. There are plenty of real examples of AI being used badly. The Deloitte incident is one. But notice what made it a problem. It wasn’t that AI was used. It was that staff trusted it blindly – and that their leaders hadn’t taken steps to prevent that from happening.
So no, I don’t think people should be shamed for using AI to write at work.
Use it. Use it often. Just remember that your writing will ultimately be judged by the same standard it always has been. Did you bring your own thinking to the table – and was the output worth the reader’s time?
As for the academic at the centre of all this: Was her article worth reading? Who knows, I never got to see it.
But that’s certainly the question I should have asked in the first place.