It’s time to address the elephant in the room; ChatGPT has a sound, a voice, and it’s beyond obvious if you know the signs to watch for. When you break down what ChatGPT is at its core, a large language model (LLM), it all starts to make more sense. LLMs are software platforms that have been fed enough data to enable them to regurgitate text back in a way that appears as if they comprehend that information. They have been programmed to predict strings of words that make the most sense in response.
The use of ChatGPT for writing content has become wildly popular. And if you look closely, you’ll see that LLM-generated content tends to say a lot of the same sorts of things… over and over.
When I see these phrases or writing styles in a piece of content (and I’ve seen them in every form, including in tweets, emails, ad reads, and even within in-person interviews with celebrities), I feel a little uneasy and disengaged. To me, the content reads as if the person or brand needed an LLM to tell them what to say. It leads me to think they didn’t actually have anything to say to begin with. So, why should I be listening?
I don’t mean to be harsh. I understand that the demands of producing content are hard to keep up with and using an LLM could mean more efficiency and output with less effort. I know that makes using an LLM desirable. But as a consumer and audience member, when I hear the ChatGPT “voice,” I immediately lose interest.
Isn’t that exactly the opposite of what we are trying to do when we communicate with people, in marketing or in general?
Three Tell-Tale Signs You Used ChatGPT to Create Your Content
Below is how I can tell someone has used ChatGPT to write their content. I’m sorry if once you’re aware of “the voice,” you won’t be able to un-hear it, but really, this needs to be said.
I also have some tips for you to bring back the humanness that audiences are craving if you’re using an LLM and notice these indicators.
1. It’s not [X], it’s [Y]
I’m putting this at the top of the list because it irks me the most. Please, please, please never use this phrasing again! This one was one of those How I Met Your Mother plate-smash moments for me. After I found out this phrase was a ChatGPTism, I couldn’t get it out of my brain. At first, I began to notice it in Instagram captions and marketing emails, but then one day I heard someone say it out loud in a TikTok I was watching. It kinda freaked me out; the woman was sitting in her car, and until that moment, I thought she was recording a fully raw, on-the-fly moment. I suppose that’s on me for forgetting that social media isn’t all real, raw moments—being in the social media management industry, I probably should have known better.
I think what frustrates me most about this type of sentence structure being overused is that it can actually be a really good way to communicate whatever point you’re making. Just like in design, contrast is clarifying, but this exact sentence is now saturating content and losing its impact as a result.
How can you avoid the trap of overusing it? Here are some suggestions:
- Simply drop the comma and give the two things you are comparing a bit more room to breathe. For example: Instead of “It’s not a harsh critique, it’s a much-needed wake-up call,” say, “It’s not a harsh critique designed to make you rethink all your life choices. Rather, it serves as a much-needed wake-up call as we navigate a new digital world.”
- Try to lose the “it’s not” or “that’s not” and just say the thing you are referring to, or other terms that reference it. For example: “This blog isn’t intended as a harsh critique; rather, I aim to provide an alternative perspective to be taken with a grain of salt.”
Literally anything but “It’s not [X], it’s [Y].” Please!
2. This = That
In a similar fashion to point #1, the “this = that” is one of ChatGPT’s favorite ways to connect ideas. I see why it appeals to marketers; it’s punchy, quick, and has that visual breakaway of the equals sign that makes content writers feel all warm and fuzzy inside. But now ChatGPT has eaten it up and taken it as its own, so if you actually want to stand out, you need to drop it.
To un-ChatGPT this one, consider using the ever-classic simile or metaphor. This may not be a muscle you’ve exercised since grade school, but we were all taught about simile and metaphor for a reason. They are effective and descriptive ways to demonstrate connection and alikeness between objects or concepts.
For example, instead of “Elevated Performance = Elevated Results”, you might instead say…”Your performance is like the muscles in your body; it needs to be exercised to increase your results.” (simile), or “Your performance is the pulley of your results; to lift something heavier, it must increase its capacity.” (metaphor)
Use similes and metaphors carefully, because sometimes they can come off as really cheesy and stupid, but with practice, they can make Instagram captions and other content more interesting and attention-grabbing. These literary devices also give your writing the human touch, setting it apart from LLM-produced content.
3. Lists With Catchy Titles
Trust me, I understand the irony of this complaint coming in the form of a list. But please, bear with me!
ChatGPT loves to make bulleted and numbered lists with each point consisting of a catchy two- or three-word title in bold, followed by a short elaboration containing one too many adjectives.
For example:
- Heat 'n Greet: Place a cozy, warm water bottle in their chilly bed for instant, comforting warmth.
Please genuinely point to someone who would naturally say the phrase, “Heat ‘n Greet.” I doubt you’ll find one.
Here’s my tip to capture the idea you want to communicate without going too far: Don’t include titles within the points of your numbered or bulleted list. Instead, take the meat and potatoes of your message and phrase it as though you are explaining it to a friend.
For instance, consider the text below as an alternative to the example above:
- Place a heated water bottle in their bed to instantly create warmth.
If you feel compelled to include titles within your listed items, try to avoid the catchy, peppy, buzzword-y vibe that every bone in your marketing body wants to use. The LLM robots love a flashy word salad, but audiences do not.
Note: You can still give the entire list a clever title to set the stage for what it covers if you’re called to do so. Just heed my advice about steering clear of trending words or phrases that have worn out their welcome.
Closing Thoughts
I fear that the more AI-generated content we consume, the more we are going to think and talk like LLMs. It’s kinda like when you’ve just finished a book or movie, and for the next few hours or even days, you find yourself thinking in the author’s voice.
In fact, even now, a little voice in the back of my head is urging me to re-read this text over and over to see if it sounds like ChatGPT.
I understand the can of worms I just opened, but it’s inevitable that we are going to have a harder and harder time proving our humanness online. It’s important for you to show your audience that you are a real person writing your own words.
Like…I’m Anna, I’m writing this at 8:43 pm, and Dawn’s gonna edit it later to be posted sometime next week. I’m a real person, and these are my real thoughts from my real brain.
There is great value in being human and writing what you actually think.
