I originally intended to write about how AI technology is progressing, but my initial research led me to an article in Marketing News Canada that gave me an “ah ha” moment that I had to share. It made so much sense that I was surprized that I hadn’t figured it out before.
The type of AI that most of us are now familiar with, like ChatGPT, are referred to as LLMs. Those initials stand for Large Language Models and mean that they have been fed huge amount of text from myriad sources to learn how to spell, write, and link different bits of information together. For the most part, what these models produce is structurally perfect. But, for me, something was always off. And now I understand.
What LLMs produce is average. Always. Average. How can it be anything else? If you take millions of books, reports, and articles, from so many different authors, and mush them all together, the result is a combination of different perspectives, tones, styles, opinions, and belief systems. What LLMs can’t yet do is connect with a specific audience from a particular point of view.
A few quotes from the Marketing News Canada article explains:
“The Canadian marketing landscape has long been relationship-driven, with trust, credibility, and reputation carrying significant weight. Those qualities are built slowly and lost quickly.”
“Voice plays a critical role here. Instead of asking whether a piece of content will perform, marketers should ask whether it sounds unmistakably like them. Voice is one of the few assets competitors cannot replicate with a prompt.”
“The brands that continue to earn attention will not be the ones publishing the most content. They will be the ones whose voice is recognizable before the logo appears.”
Brands need to be distinctive, and need to appeal to their particular market. They need to be anything but average.
Here’s a link to the full article: Long Live the Typo: How to Navigate AI Fatige in Modern Marketing
