How to remove AI tells from your writing

"AI tells" are the recurring fingerprints that make text read as machine-written. Most come from a simple fact: language models pick the most statistically likely next word, which pushes writing toward smooth, generic, oddly uniform prose. Here's the practical checklist, grouped by type, with the fix for each.

Vocabulary tells

Certain words are wildly over-represented in AI output. When you see them, replace them with plainer language.

• delve, testament, tapestry, landscape, realm, leverage, utilize, myriad, plethora, intricate, seamless, robust, pivotal, foster, bolster, embark → use, key, rich, smooth, strong, start.
• "a testament to" → show the evidence directly.
• "plays a crucial role" → say what it actually does.

Cadence tells

• Negative parallelism — "It's not just X, it's Y." State the point directly.
• The rule of three — "innovation, inspiration, and insight." Keep only the items that earn their place.
• Manufactured punchlines — "No preference. No prior. No nostalgia." Vary sentence length instead of stacking fragments.
• Aphorism formulas — "Symmetry is the language of trust." Replace with the real claim.

Filler & hedging

• "in order to" → "to"; "due to the fact that" → "because".
• "it is important to note that", "needless to say" → cut entirely.
• Stacked hedges — "could potentially possibly" → "may".

Structure tells

• Em-dash overuse — swap for commas, periods, or parentheses.
• Emoji in body copy — remove.
• Bold overuse and Title Case Headings — reserve bold for real emphasis; use sentence case.
• Curly quotes straight from a chatbot — normalise them.

Chatbot leftovers

• "Great question!", "I hope this helps!", "Let me know if…" — delete.
• Cutoff disclaimers — "As of my last update…" — remove or replace with a real source.
• Generic conclusions — "The future looks bright." End on a concrete fact or next step.

The one that isn't a word: rhythm

Even after you fix the words, uniform sentence length gives writing away. Humans write with burstiness — short punchy sentences next to long winding ones. Read your draft aloud; if every sentence is the same length, break some up and let others run.

Roughly two-thirds of these patterns can be caught with plain rules — which is why Naturable's free checker can flag and fix them instantly, with no AI involved.

Related: What are perplexity and burstiness? · Do AI humanizers actually work?