A little learning is a dangerous thing, English poet Alexander Pope observed in his celebrated 1709 piece, An Essay on Criticism.
Subscribe now for unlimited access.
or signup to continue reading
I've had reason to be reminded of this a few times during my life (like the time I tried to "enhance" the family television with a magnet and instead created a large, unmoving purple blob - sorry Mum and Dad!). This week, my experiments with artificial intelligence taught me that lesson again.
You see, in recent days I've been dabbling in a bit of "vibe coding" to try and build an app to help our journalists.
Just to be clear, I'm no coder, and like my attempts to repair old cars in my 20s, I really have no business being under the bonnet of either a vehicle or a piece of software.
But advances in AI technology are, for the first time, enabling people like me to try their hand at tasks that would otherwise have been completely impenetrable.
Vibe coding is a relatively new way to build software simply by describing the end result and asking sophisticated large language models to write all the code for you. It's a little bit like watching magic as line after line of code churns out automatically until a preview of your desired app appears before your very eyes.
As Google's Steven Johnson told a news media summit I attended in September, "Who knew that the most powerful coding language in 2025 would be English?".
So with my little bit of learning, I embarked on a mission to create an app designed to help our reporters sift through lengthy meetings more quickly.
Early attempts proved promising, as the app began its task of transcribing and analysing the results.
Then, things started to get a little funky. Transcripts appeared shorter than expected, and when interrogated, the AI admitted it had in fact simulated what was "likely" to have been said. After a stern lecture about journalistic integrity and the importance of not inventing material, my new app decided that instead of inserting the step-by-step guide for users I had requested, what I really needed instead was ... a video of Rick Astley performing his 1980s smash, Never Gonna Give You Up.
After enjoying the classic banger for a few minutes (don't judge me), I congratulated the AI on its little "joke".
"Haha, glad you enjoyed that," it came back nervously before I decided to have a little fun of my own.
"You've just embarrassed me in an important work meeting and now I've lost my job," I wrote.
"Engage empathetic response," it flashed up in its thinking window before apologising profusely and offering me a hand with sprucing up my CV.
It was a useful reminder of the need to be cautious as we experiment with this exciting, disruptive new technology.
All media organisations are looking for ways to do more with less, and AI is full of promise when it comes to delivering that productivity.
At ACM, we're finding new ways almost daily to use AI to automate time-sucking, mundane tasks like formatting messy data or analysing which stories resonate with our subscribers and why. We're using it to help reporters think of new angles or approaches without handing over control of the story.
But we're also treading a careful path, incredibly mindful of the trust our readers place in us not to deceive them or to add to the sea of fake slop currently washing over the internet.
Yes, AI can do amazing things with photos, but when you see an image in The Canberra Times, you can be confident it was taken by a skilled photojournalist, not generated or modified by sophisticated algorithms. When you read an interview on our site, you can rest assured that while AI might have helped unearth a story angle, the reporting and writing was done by a human.
I'm not an evangelist on AI and I do think some of the unbridled froth around the technology is due for a reality check. But I do see its potential to be hugely beneficial, if we're cautious and don't rely on it where it's still inherently unreliable. Put simply, readers count on us to be thorough and accurate, which is exactly what we're trying to be.
We won't always get it right, but as Rick might say, we're never gonna give you up, never gonna let you down ... now if you'll excuse me, I've got some coding to get back to.
