This matters because
As a consumer, I love journalism. There is no finer satisfaction in life than reading the well written words of a professional writer, who is speaking on the big subjects of the day. But that love is starting to feel dirty and it’s not the fault of the journalists.
As a teenager I was a vociferous consumer of print media. Every weekend brought a stack of Sundays that I would dutifully work my way through. In my late teens, every day meant a new Liverpool Echo and every Wednesday a copy each of the NME and Melody Maker, both smelling of newsprint and paper.
I grew up loving journalism and to this day, I love to read. It’s not doom scrolling for me, it’s time well spent understanding the world and its people.
But a couple of things have been killing this love for me lately. The first of them is the phrase ‘this matters because’.
It’s the hallmark of the lazy writer, who has ‘borrowed’ most of the words in which you are investing your time directly from ChatGPT.
My other big bugbear is when an article contains a claim followed by a statement of what that claim’s subject isn’t. For instance, “Fashion in 2026 is less about trends dictating taste and more about signals of identity, intent and awareness.”
Or “Industrial robotics in 2026 is about adaptability, integration and usable intelligence, not just hardware. It’s not just about machines doing tasks to systems delivering outcomes.”
In AI language this horrible writing style is referred to as a ‘structural marker of reasoning’. Essentially, it’s a word or phrase that either precedes actual reasoning or, in many cases, replaces actual reasoning.
You will notice that both the fashion and industrial robotics examples above are almost entirely meaningless, but they both sound like they mean something until you actually pay attention. Both were generated by ChatGPT.
As well as often being utter nonsense, ‘either or’ reasoning only works in ladder logic; and humans don’t run on ladder logic. Normally, it’s not clearly defined alternatives that are interesting to us; we are more into the grey areas in between.
Why must we suffer this SPAM?
AI generated articles do this because LLMs (Large Language Models) like structured answers to clear questions. It helps them to deliver the type of information they think we want in answer to our questions.
Unfortunately, humans don’t always like this kind of A equals B structure in writing. We find it a bit irksome and obvious; it lacks poetry and soul. It makes you feel a bit dirty when you read enough of it and it doesn’t reach us on a meaningful level.
Given that most of the people you want to influence are human beings, it makes sense to avoid using the phrases and giveaways in your writing that makes it look like it’s been written by an AI. Journalists get hundreds, sometimes thousands, of emails everyday and if yours smells like bad AI, they aren’t likely to hang around long enough to give you the benefit of the doubt.
So, don’t rely on an LLM to write content for your PR and marketing campaigns; rely on a human. I don’t mean don’t use AI at all; I’m a huge, huge fan. Just don’t replace yourself with it. After all, you don’t want your readers left feeling dirty.
We get that it’s hard to keep up with the latest thinking in PR and marketing. Subscribe to our twice weekly blog, direct to your inbox, for the answers to all the burning questions your board ask – from using AI in marketing to how to spend your budget against a backdrop of constantly multiplying channels.
