18 Secret Indicators of AI Writing (And the Exact Prompt That Fixes All of Them)
You've been reading AI-generated content without realizing it. Here's how to spot it and stop writing like a robot yourself.
You've seen these patterns everywhere.
In captions. In blog posts. In LinkedIn essays that feel weirdly polished but say absolutely nothing.
You just didn't know what to call them.
AI writing has specific fingerprints. 18 of them. And once you see them, you can't unsee them.
I'm going to break down every single one. With real examples. So you can spot them instantly, whether you're reading someone else's content or checking your own.
Because here's the thing that should bother you: a lot of people have started writing like this themselves. Without even realizing it. We're absorbing AI language patterns just by consuming AI-generated content every day.
Let's fix that.
Sentence Structure Patterns
1. Negative Parallelism
The "It's not X, it's Y" pattern. Before LLMs, people simply did not write like this at scale. It includes the causal variant "not because X, but because Y" where every explanation is framed as a surprise reveal.
Examples:
"This isn't just technical, it's a mindset change."
"The question isn't whether to optimize. The question is when to stop."
"Half the bugs you chase aren't in your code. They're in your head."
2. Not X. Not Y. Just Z.
The dramatic countdown pattern. AI builds tension by negating two or more things before revealing the actual point. Creates a false sense of narrowing down to the truth.
Examples:
"Not ambition. Not ego. Just a person who never learned to say enough."
"Not broken. Not behind. Just building something no one has a name for yet."
3. False Ranges
Using "from X to Y" constructions where X and Y aren't on any real scale. AI uses it as a fancy way to list two loosely related things. "From innovation to cultural transformation" ... what's in between? Nothing.
Examples:
"From innovation to implementation to cultural transformation."
"From fundamental physics to medicine and neuroscience."
4. The X? A Y.
Self-posed rhetorical questions answered immediately. The model asks a question nobody was asking, then answers it for dramatic effect.
Examples:
"The result? Devastating."
"The worst part? Nobody saw it coming."
"The impact? Immediate."
5. It's Worth Noting
Filler transitions that signal nothing. AI uses these to introduce new points without connecting them to the previous argument. Also includes: "It bears mentioning", "Importantly", "Interestingly", "Notably".
6. Tricolon Abuse
Overuse of the rule-of-three pattern. A single tricolon is elegant. Three back-to-back tricolons are a pattern recognition failure.
Example: "Products impress people; platforms empower them. Products solve problems; platforms create worlds."
7. Anaphora Abuse
Repeating the same sentence opening multiple times in quick succession.
Example: "They assume that users will pay... They assume that developers will build... They assume that ecosystems will emerge..."
8. Superficial Analyses
Tacking a present participle ("-ing") phrase onto the end of a sentence to inject shallow analysis. Things like "highlighting its importance" or "reflecting broader trends" or "contributing to the region's rich cultural heritage."
Tone Patterns
9. Grandiose Stakes Inflation
Everything is the most important thing ever. A blog post about API pricing becomes a meditation on the fate of civilization.
Examples:
"This will fundamentally reshape how we think about everything."
"You are not running a race. You are setting the pace for everyone behind you."
10. Invented Concept Labels
AI appends abstract problem-nouns (paradox, trap, creep, divide) to domain words and uses them as if they're real, established terms. "The supervision paradox." "The acceleration trap." "Workload creep." Name a thing, skip the argument.
11. Here's the Kicker
False suspense transitions that promise a revelation but deliver a point that didn't need the buildup.
12. Think of It As...
The patronizing analogy. AI defaults to teacher mode and assumes you need a metaphor to understand anything. Often the analogy is less clear than the original concept.
13. The Truth Is Simple
Asserting that something is obvious instead of actually proving it. If you have to tell the reader your point is clear, it very likely isn't.
14. Vague Attributions
Attributing claims to unnamed authorities. "Experts argue..." "Industry reports suggest..." If you can't name the expert, you don't have a source.
Word Choice, Paragraph Structure, and Formatting
15. Tapestry and Landscape
Overuse of ornate nouns where simpler words would do. "The rich tapestry of human experience..." "Navigating the complex landscape of modern AI..." Other offenders: paradigm, synergy, ecosystem, framework.
16. Short Punchy Fragments
Excessive use of very short sentences as standalone paragraphs. "He published this. Openly. In a book. As a priest." RLHF training pushed models toward one thought per sentence. No real person writes first drafts this way.
17. Listicle in a Trench Coat
Numbered points dressed up as prose. "The first step is research. The second is drafting. The third is review." It's a listicle pretending to be an essay.
18. Unicode Decoration
Unicode arrows, smart quotes, and special characters you can't easily type. Real writers produce straight quotes and --> or =>. Claude especially loves the arrow character.
The Fix: One Prompt That Handles All of This
I wrote a single prompt you can paste into ChatGPT, Claude, or any AI before you write anything. It tells the AI to avoid every single one of these 18 patterns. I use it myself every day.
Here it is. Copy the whole thing:
Write in a natural, human voice. Avoid these patterns: no 'It's not X, it's Y' constructions, no em-dashes, no 'Not X. Not Y. Just Z' countdowns, no fake 'from X to Y' ranges, no inflating stakes to world-historical significance, no self-posed rhetorical questions answered immediately, no filler transitions like 'It's worth noting' or 'Importantly', no invented compound labels like 'the supervision paradox', no ornate words like 'tapestry' or 'landscape' where simple words work, no tricolon or anaphora abuse, no excessive short punchy fragments as standalone paragraphs, no 'Here's the kicker' false suspense, no patronizing 'Think of it as' analogies, no asserting something is simple instead of proving it, no listicles disguised as prose, no superficial -ing analysis tacked onto sentences, no vague 'Experts say' attributions, no unicode arrows or smart quotes. Write like a real person thinking out loud. Vary your sentence length naturally. Let some thoughts be messy. Not everything needs a punchline.
That's it. Paste it at the top of any conversation where you're asking AI to write for you.
It won't make AI writing perfect. But it removes the most obvious tells. The stuff that makes people scroll past your content because their brain pattern-matches it as "AI slop" in half a second.
Your audience deserves better. And so does your voice.
Original research sourced from tropes.fyi and Wikipedia's Signs of AI Writing page. Shoutout to @ameerahbadr for the original carousel that inspired this breakdown.