Move over, em dash. There’s a new AI tell in town.
We’ve all learned about certain words, language patterns, and punctuation that AI tools like ChatGPT use over and over. Here’s one I’ve started seeing a lot: adverbs like honestly, personally, quietly, and genuinely.
In ChatGPT responses, these words usually don’t add factual information.
Instead, they indicate how you, the reader, should feel about or interpret something. They nudge you to consider the response as being more candid, thoughtful, personal, or important.
In linguistic terms, these words are acting as stance adverbials (adverbs that tell readers how to interpret a statement) or discourse markers (words that signal how one idea relates to the next).
Because they’re more about vibes than facts, these words can imply characteristics an LLM doesn’t actually have, like personal experience, genuine feelings, or an inner point of view. Things can get weird when you stop to think about what it means for an LLM to speak “honestly” or “personally.”
I’ve encountered a bunch of these when working with ChatGPT recently. Here are a few examples from my own chats:
Honestly
What does it mean for ChatGPT to be “honest” with you? Does that imply it’s sometimes not being honest? In AI responses, “honestly” is usually just a rhetorical cue that advice or a recommendation is coming next:
“Honestly, Stephanie, I think this is your answer.”
“Honestly, I think this is a pretty sophisticated growth edge for you.”
“Honestly” adds warmth without adding information. It can make the conversation seem more candid and personal, as though there’s an expert behind the scenes giving you advice from experience. Of course, ChatGPT is neither an expert nor a person with lived experience. It’s simply a language model generating the next most likely words.
That’s why I find its use of “honestly” to be a little jarring.
Personally
People use “personally” to distinguish their own opinions, preferences, or experiences from objective facts. Of course, ChatGPT doesn’t have any of those. But it still sprinkles “personally” liberally throughout its responses:
“Personally, I think the real design flaw isn’t the per-device setting.”
“Personally, I’d standardize to…”
“Personally, I’d pause on how you actually use this material.”
The word makes advice sound more conversational, as if it’s coming from a human advisor. Now ChatGPT sounds like a thoughtful colleague sharing a personal opinion.
Even if it’s just a rhetorical device, it gives me the sentient computer heebie-jeebies. (Like HAL 9000: “I’m sorry Dave, I’m afraid I can’t do that.”)
Genuinely
Sometimes “genuinely” is just an amplifier that could be replaced with “very” or “really.” But it sounds warmer and more sincere.
“This one is genuinely confusing because…”
“I’d actually lean into something you genuinely enjoyed from the conversation…”
The word suggests authenticity, as if ChatGPT is saying, “Believe me, I really mean this.” That’s a natural way for people to talk, but ChatGPT doesn’t have genuine feelings, preferences, or convictions to express.
Like “honestly” and “personally,” this one sounds weird to me coming from ChatGPT.
Quietly
Good gracious, can you get a response from ChatGPT these days without at least one “quietly” in it? It seems to love this word.
“Those are all examples of people quietly voting with their feet.”
“The risk with highly operational contractor roles is that they can quietly narrow your narrative over time.”
“Quietly” suggests that ChatGPT is revealing a trend or change that isn’t immediately obvious. It’s giving you insider knowledge or subtle insight that other people haven’t realized yet.
Don’t you feel smart? ChatGPT would like you to think so.
Simply
Sometimes this adverb is just there for emphasis, and it’s not doing much work. More interesting, though, is when ChatGPT uses “simply” to suggest that a claim is obvious or self-evident. Or it can make a statement sound more certain than it really is:
“…their goal is simply to verify the existence of the Schedule C.”
“At this point, I’d simply keep the money in the savings account.”
Instead of supporting the claim with evidence, it nudges you to accept the conclusion without further explanation.
Actually
“Actually” has a perfectly legitimate job: it signals a correction, contrast, or surprise. But ChatGPT often uses it even when there’s nothing to correct:
“Do we have all the stuff needed to actually put this in front of customers?” (Why wouldn’t we?)
“This is actually interesting.” (Did anyone say it wasn’t?)
ChatGPT is giving you the subtle nudge that you’re supposed to find something unexpectedly true or important, whether it’s justified or not.
Unusually
“Unusually” suggests that something stands out from the norm. I’ve noticed that ChatGPT uses this word most often in compliments. It praises you by implying that you’ve exceeded some invisible benchmark:
“Where you’re unusually strong…”
“Your current checklist already reflects unusually good instincts.”
With “unusually,” ChatGPT gives the impression that it has carefully calibrated your performance against a broader standard, though the comparison isn’t made explicit.
Bonus: Lately and recently
This one isn’t really about the adverbs themselves. It’s about the way ChatGPT uses them as part of a larger rhetorical pattern:
“This is one of the more promising fits you’ve shown me lately…”
“This is one of the more interesting jobs you’ve shown me recently…”
While I was using ChatGPT to analyze job postings, these statements suggested it was comparing and ranking everything we’d discussed. But I noticed that it said some version of this about almost every job posting I shared.
So why does it use this phrasing if it doesn’t mean anything? I think it’s meant to make the conversation feel continuous and collaborative. It reminds you that ChatGPT remembers what you’ve been discussing and reinforces the idea that you’re working together.
Despite what it may suggest, this phrasing wasn’t there to make a helpful comparison, but to make the conversation feel more personal and engaging. To me, it feels like insincere flattery.
So do I need to stop using these adverbs?
None of these words are inherently bad or manipulative. They all have legitimate uses, and skilled writers use them all the time. Sometimes “actually” is actually the right word. Sometimes “quietly” captures a gradual change perfectly.
The problem is that ChatGPT uses these adverbs over and over, much more often than you’d see them in regular prose. It has become so fond of certain words and phrases that they’re starting to sound formulaic.
I use ChatGPT so often that I’ve developed a knee-jerk reaction to some of them, particularly “quietly” and “genuinely.” Even when you use them intentionally, readers may subconsciously associate them with AI-generated writing. Thanks a lot, AI.
Strunk and White famously wrote, “A sentence should contain no unnecessary words.” Before you use one of these adverbs, ask yourself: What work is this word doing? Is it adding information? Clarifying your meaning? Creating a legitimate contrast? If not, try taking it out. You might find that the meaning is the same, but your sentence is crisper.
As always, it comes back to using your judgment and not outsourcing your editorial chops to AI. Tools like ChatGPT are powerful writing partners, but it’s still your job to decide which words belong and which ones don’t.