Are We All Starting to Sound Like ChatGPT?

A few years ago, if you told someone their emails “delved” into a topic or called a process “meticulous,” you might have sounded a little stiff. Now, words like these have crept into everyday conversations, LinkedIn posts, even YouTube lectures. Strangely enough, the reason might be the same technology many of us use every day: AI chatbots like ChatGPT.

More Than Just a Tech Trend

When ChatGPT first launched in late 2022, most people saw it as a handy assistant: a quick way to summarize a memo, find the right words for an awkward email, or polish up a report. But according to several recent studies, the effect has gone far beyond convenience. Researchers at the Max Planck Institute, for example, reviewed hundreds of thousands of academic YouTube videos and found that after ChatGPT became widely used, certain words—“meticulous,” “delve,” “realm,” “adept”—started popping up everywhere, sometimes over 50% more often than before.

It’s not just academic videos, either... Similar patterns are showing up on Twitter (now X), in emails, and even in casual online chats. Linguists are calling it a “linguistic watermark”—a subtle sign that AI is leaving its stamp on our vocabulary.

The Subtle Shift in How We Communicate

It’s not just the words themselves... AI, it seems, is also influencing the rhythm and structure of our sentences. Instead of short, punchy statements, we’re seeing longer, more formal sentences, with a bit of a neutral, “polite” tone—just like the ones you’d get from ChatGPT. Even the way people speak is starting to shift. New research suggests that when people interact with voice-based AI, they unconsciously begin to match the bot’s cadence, pacing, and even emotional flatness.

In a way, it’s like picking up an accent after spending a few months in a different city. But instead of Paris or Rome, we’re all picking up the ChatGPT accent.

The Double-Edged Sword of AI-Polish

Is this a bad thing? On one hand, these changes can make communication clearer, more organized, and less abrasive. Some researchers at Cornell even found that AI-written responses help foster cooperation and positivity. But there’s a catch: when people suspect they’re being responded to by a chatbot, they tend to trust less, feeling that something genuine is missing.

Language is more than just words and grammar. The little quirks—the awkward jokes, the typos, the imperfect metaphors—are what signal humanity and effort. As AI becomes more prevalent, experts like Professor Mor Naaman warn that we risk losing those signals. Conversations might sound polished, but they could also feel impersonal, bland, and ultimately less trustworthy.

Why It Matters for Culture and Creativity

There’s a growing concern among linguists and educators that as AI-driven phrasing spreads, it might slowly erode cultural quirks and local expressions. A recent Atlantic article described this as the “Great Language Flattening”—a world where everyone, everywhere, sounds the same because everyone is copying from the same source.

Think about regional slang, inside jokes, and the little ways we show who we are through language. If those fade away, we lose some of what makes human connection unpredictable—and, frankly, fun.

What Should We Do?

The easy advice is: “Be aware,” or “Use AI mindfully.” But that barely scratches the surface. If AI is shaping how we write, speak, and even think, then the challenge isn’t just individual—it’s social, even cultural. We need to think bigger.

First, we should start treating language like a shared resource, not a neutral tool. The way we phrase ideas, the tones we adopt, the stories we tell—these shape how societies function. When AI smooths out our rough edges and flattens our expressions into polished, sterile phrases, we lose more than flavor—we risk losing friction, emotion, the spark of dissent. So the question becomes: What kind of language do we want to circulate in public life?

Second, it may be time to create new norms for hybrid expression. Instead of either “human-only” or “AI-enhanced,” we might evolve new genres of communication where people intentionally weave in their imperfections. Think of it like jazz: a deliberate off-beat, a bent note, a pause in the wrong place—not mistakes, but style.Could you end an AI-polished memo with a sentence you typed without editing? Add one line you wrote quickly, from your gut? These gestures may seem small, but they reassert presence. They say: a person was here.

Third—and this may sound odd—we need to teach resistance to linguistic mimicry. Just like media literacy, where we teach people how to spot bias or misinformation, we may need a kind of language literacy: learning to detect when your own tone has drifted into autopilot. Not because that tone is “bad,” but because it might not be yours. Your voice might be hiding behind a GPT-friendly script.

Lastly... there’s room for play. We could start collecting endangered words, celebrating strange metaphors, encouraging local dialects in global spaces. Maybe writers, educators, and designers need to push harder against the smoothness of AI prose and champion language that disrupts, that feels a bit “off,” that makes you pause. Because that’s where thought begins—not in what flows, but in what stops you.

Let’s Not Forget: We’re Shaping AI, Too

When people talk about how AI is reshaping our language—making it smoother, more formal, more predictable—they often miss the more complicated truth: AI is mimicking us. Every email we write, every forum post, every polite phrasing we let autocorrect glide into our messages—all of it becomes training data. The “ChatGPT tone” isn’t just something imposed on us. It’s a mirror of what we’re already doing.

If that’s the case, the real question isn’t just how to resist AI influence, but what kind of influence we want to give in return.

For example... Curate the Feed, Don’t Just Consume It.

Every time you rewrite a rough sentence to sound more like a template, you’re contributing to the erosion of uniqueness. But when you leave in a strange word, a human pause, a phrase that doesn’t quite “fit,” you’re keeping something alive. Use the AI, but leave your fingerprints.

______

FOLLOW US FOR MORE.

DATA STUDIOS

datastudios.org