Ep. 34: 8 AI Writing Patterns That Are Dead Giveaways
For the love, please stop using AI to write your copy. We can all tell. In this episode I call out the 8 most obvious AI writing patterns, and provide specific names for those specific styles that are just so characteristically robotic. From em dashes and antithesis structures to hollow opens and metronomic cadence, I break down why these patterns appear, why AI defaults to them, and why the best solution is usually just writing it yourself.
Scratching the Itch
Why this episode? Because sometimes it’s really nice to be able to put words to things that are so readily apparent but have gone nameless. That’s literally it. I want to give you language for the things that I know you 100% see and feel when you’re reading Threads and captions and emails and you’re just like: This was absolutely written by AI.
My goal isn’t to give you prompts to make AI sound less…AI. I literally just want to give you words for those immediately recognizable AI writing patterns. Why? Because having words for things makes my brain feel better, like when IMDB tells you the name of the obscure movie that the supporting actress on your tv screen was in 14 years ago. It just scratches an itch that feels so good, at least it does for me, and I’m hoping it’ll do the same for you.
Alright, let’s dive in.
The 8 Patterns
1. Antithesis Structure: “It’s not X, it’s Y.”
Lawd jaysus, these drive me crazy. People don’t actually speak like this in real life, which makes it feel hugely performative…which is exactly what it is.
2. Em Dashes
JFC. The em dashes. I feel bad for anyone who used them before AI was a thing. I was not one of those people, and if you haven’t yet noticed, I will use the shit out of parentheses and commas. Never been one for an em dash or a semicolon. If your writing suddenly has em dashes everywhere and it didn’t before 2023, people gonna notice.
3. Parallel Structure Overload
This is a BIG one. Everything comes in threes: it’s balanced, it’s symmetrical, it’s giving robotic. Why is it so jarring? Because real writing is uneven…because thinking is uneven.
4. Metronomic Cadence
Speaking of the unevenness that is naturally human: human writing speeds up, slows down, breaks off. AI, however, maintains a measured pace throughout. Nothing is abrupt, nothing lingers, and there is little to no variation in structure. It reads like a metronome and it feels like a robot…because it is.
5. The Illusion of Consideration
These ones drive me CRAZY and are, IMO, the new telltale kid on the block. Em dashes were the first easy tell, then it was the antithesis structure. Now you’ll see, especially on things like Threads: “Here’s what they don’t tell you.” Or, “This is where most people get it wrong.” It’s actually the worst. It sets up a reveal and delivers…nothing.
6. Adjective Stacking
It’s giving keyword stuffing. Ex: “Robust, scalable solution.” “Thoughtful, nuanced approach.” The adjectives are always complementary and always vague.
7. No Risky Sentences
Everything is predictable and fits the pattern. It’s like when you’re watching a poorly written movie and you can guess exactly what the character is going to say despite the fact that you’ve never seen the movie before. The dialogue is formulaic and there are no genuine opinions…because there’s no human on the other end to generate them. This is what happens when writing is produced by math, instead of from the heart.
8. Hollow Opens and Tidy Closes
Responses generated by AI start with context-setting and end with a bow. Every opening is a metaphorical robotic throat clearing that just wastes your time and your tokens. Worth noting: You can give your AI instructions to not do this and it will generally respect those parameters decently well, but you’ll likely always still get some throat clearing. Real writing can start and stop anywhere, and is largely dependent on the individual’s writing style. I know that for me personally, I end my paragraphs (and my writing in general) when the feeling is resolved, not when the pattern requires a period.
Why AI Writes Like This
So, why does AI write like this? :::Insert drumroll::: Because AI is a computer program that was trained on human writing and now simply does math to generate the answer. (We’ve been talking about this since episode 1!)
Yes, real talk, the human brain is also doing this at a much faster rate and higher level, don’t get me wrong. But we have free will, sentience, we are self-aware, and we have a soul. This whole discussion about sentience and being self-aware honestly fascinates me and is definitely a topic I want to dive deeper into. Unfortunately I currently only have 24 hours in a day, so I don’t have time for that deep dive right now, but when I do it’s gonna require learning from a bunch of different fields: cognitive science, philosophy, neuroscience, psychology…and most notably brown people and women.
But for now: AI writes like this because it was trained (yes, largely illegally) on basically all the text and writing that exists on the internet, and those patterns that I just listed out appear a lot, typically in a positive light. So AI reproduces them.
Statistically speaking, they “look” like “good” writing.
If you’re sitting there like, “WTF kind of good writing is it pulling from, because I’ve never seen any of that nonsense in the stuff I’ve read.”…you’re right. The reality is, your reality (aka, the fact that you have good taste) does not reflect EVERYTHING that exists on the internet. AI is trained on EVERYTHING, not just the actual good stuff. Think about all the whack ass self-help IG accounts out there with millions of followers. Yup, it’s trained on that nonsense and those captions.
So What Can You Do About It?
While getting AI to sound less AI-like is absolutely not the intention of this episode, I want to throw in this section so I can speak to the bigger picture.
Is there anything you can do to have your preferred AI sound less like AI when it generates writing? Yes and no. You can prompt it, you can give it your writing samples, you can provide instructions to avoid certain patterns. But you’re always going to be fighting against its underlying tendency. This is why editing is ALWAYS helpful, and honestly your best bet is to just simply write it yourself.
FWIW, I understand the appeal of having AI write things for you. Particularly during the times where you just want that piece of work to be done. There’s also that slot machine appeal dopamine hit of “what might this thing generate?!”. But, as I’m sure you’ve experienced, if you have to correct AI enough times, eventually you just say, “Screw it, I’ll do it myself.” Very much akin to continuously losing at a specific slot machine in Vegas and finally getting fed up and walking away.
Again, my goal in sharing these 8 telltale AI writing patterns was simply to give you words for what you’re seeing. If you want to use the list to audit your own AI-generated content, so be it. But, hopefully this list has shown you that there is a lot that will need to be edited if you’re having AI write for you…so you might as well write it yourself.
How I Used Claude This Week
Each episode I share a quick example of how I used AI that week.
This week I used it to very quickly fix the background color AND link color on a page on my website. I’ve talked many times about using AI to generate code, but what I want to highlight in this instance is that I knew exactly what to ask for. I knew I needed CSS and I knew how to inspect the page to be able to target the element that I needed that CSS for.
So how is this use case any different than what I’ve spoken about before? Because it demonstrates that over time I’ve used AI to learn about web development, in this case more specifically, front-end development. So, yes, folks can look to argue that AI makes you lazy or dumber, and I won’t argue with them…because I don’t argue with anyone. But I will use my own platform to stand 10 toes down on the fact that AI can also absolutely help you learn things you very likely otherwise wouldn’t have learned.
Da Wrap-up
There you have it, 8 AI writing patterns that are dead giveaways. Hopefully I scratched the naming itch, and perhaps even more than that, convinced you to just write the things yourself. It’ll sound better.
As always, endlessly appreciative for you and your curiosity.
Catch you next Thursday.
Maestro out.