This article caught my attention over the weekend. AI is pretty much everywhere, and at least for now still fairly easy to 'spot' among non-AI generated content. The point this article makes, that AI has a newer, fairly obvious language pattern, resonated because not only have I noticed this, but I've caught myself questioning some of my own writing and whether it's starting to take on any of that tone. I'm sure I'm not alone.
In short, the article points out that AI has a few distinct language patterns. We're all aware of the Em dash—which I've always personally loved—but now need to be very judicious about its use. Another is more of a pattern in the language itself, and makes me think about how pacing and tonality, not just the words and letters and punctuation themselves, create our experience of language, and how AI may actually be changing that experience in real-time. AI-generated content has what the article refers to as negative parallelism. Describing a thought "NOT AS X", "BUT AS YZ". We've all seen it by now: "It's not happening in disparate moments—it's here to stay".
What I think is happening with this "negative parallelism" is that AI is at its core a reasoning tool. It wants to find the most logical output or answer. It's excellent at analyzing and applying patterns to find the most logical path forward. And so, with this AI language tic, I think what it's doing is reasoning in real-time, calling out what something is NOT as a way to point the reader to what must be true, almost as a way to cover it's own bases. In strategy, we work in something called Structured problem solving, which is really just a framework for finding a hypothesis and attaching any data that, when combined, points you to the hypothesis as the logical answer. The scientific method does the same thing. As we know, science, strategy, and, yes, language are all nuanced. But in the face of unlimited paths forward, the strategist or scientist or AI must find a way to prove their chosen path. The article points out a few reasons this tic might be happening, such as human content actually training it this way; or that it's trained to find the path of least resistance to back up a point (doubling down on its output), which are probably also at play. But I think that as a logistician, the AI is coming to a conclusion that when consumed, leads the reader to believe it must be true, introducing a polarity that most of us naturally filter in our thought process without needing to verbalize unless we are making an important point.
This, at least for me, is creating anxiety and tension as I consume AI content. The polarity introduced by saying "not only X, but Y.." creates an urgency around everything which human reasoning and language would normally filter through nuance. This new AI language doesn't leave much room for nuance, and with such definitive speech patterns, we should remember that very few topics are as urgent or as polarizing as every output generated digitally these days. Not a criticism, just an observation.
I'm currently working on a project that's close to my heart, one that isn't finished or ready for showing yet, but which highlights the reverence around craft, tradition and the makers and authors behind it. I'm not sure if it's part of growing and getting older, but I've been drawn lately to anything with deep history, tradition, and hand-made qualities: My family has German roots, and there are particular traditions we hold close such as crochet and-yes-sourdough. But also things such as the quality of sound coming from records versus streaming, song writing, real, authentic style whether thats in home decor, fashion, or art, and all of the nuances and mistakes and happy divergences that go along with these experiences.
Perhaps one of the most obvious impacts that AI has already had has been on the written word, on language, and this article reminded me of that while raising an interesting question: What is the new language of AI creating long-term? I'm not talking about what it means for data, intelligence, collective reasoning.... I'm talking about watching in real-time the rise of a new language. A way of combining words and sounds and punctuation. It hadn't occurred to me, but I'm realizing how much language actually impacts our experience of a thought, or of how to reason. A language that's more descriptive and beautiful imparts a different tone than one which is curt, linear. One isn't better or worse than the other, but they convey different feelings and inherently demand different pacing of information. AI is excellent and getting a point across, doubling down, and moving to the next thing. However, the magic of the journey—what's captured in the awkward words or accidental brush strokes in a painting—is missing. And it's being trained on ITSELF, so naturally over time, the nuances will continue to filter out. I'm curious as to how AI developers will manage this, or if they even will.
I hadn't considered this until recently, but we really are watching the rise of a new language in real-time, cutting across cultures and geography, and since language is the core of a culture, I wonder how this will shape the collective us moving forward: Will we stop writing? Will we stop going through the clunky process of coming to a conclusion, a song bridge, a plot twist? And most importantly, if we do slow down our original works in progress, how will AI continue to interpret our history up until when the written word slows down? Or will it create traditionalists, crafters, artisans who double-down on their work? Is the rise of AI language creating even more polarity, or is it pushing us to further explore our own reasoning and originality, and embrace the messy, subtle middle...time will tell, and in the meantime I have a nuanced project to get back to..