AI-ism Checker
Paste any text and get counts of em dashes and the phrases that make writing read as AI-generated, like delve and it's not just.
Em dashes: 0
Paste text above to check it.
This is a heuristic, not a detector. Plenty of honest human writing uses these phrases, and a carefully edited AI draft can score zero. Treat a high score as a prompt to reread, not a verdict.
How it works
Paste in a paragraph and the checker runs it against two things: a short list of words and phrases that show up constantly in AI drafts, and the raw count of em dashes. Both are counted, then added together into one flag total, along with word count and average sentence length so you can see the shape of the writing at a glance.
Worked example: the sentence "It's not just fast, it's a game-changer, truly" written with two em dashes instead of commas, "It's not just fast — it's a game-changer — truly.", scores 4. Two of those points come from the em dashes themselves. The other two come from matching "it's not just" (the classic negative-parallelism setup) and "game-changer" (a stock AI intensifier) once each. Swap the dashes for commas and cut the two phrases, and the score drops to zero without losing the point of the sentence.
FAQ
Does a high score mean the text was written by AI?
No. It means the text contains patterns that show up more often in AI writing than in typical human writing. Some people genuinely write this way, and a well-edited AI draft can dodge every phrase on the list. Use the score as a place to start editing, not as proof of anything.
Why count em dashes separately from the phrase list?
Em-dash density is, on its own, one of the strongest single tells we've seen in AI drafts. Counting it apart from the phrase matches lets you see whether a high score is coming from punctuation habits, word choice, or both, which changes how you'd edit the piece.
What phrases does it look for?
Words like delve, unlock, elevate, and tapestry, plus phrasing patterns like "it's not just X, it's Y", "whether you're a beginner or a pro", and "in today's fast-paced world". These are the phrases that show up so often in AI output that readers have started to recognize them on sight.
Can human writing trip this checker?
Yes, and that's expected. Someone who genuinely likes the word delve or leans on em dashes will score high here even if every sentence is their own. The tool flags patterns, not authorship. Read the flagged spots and decide for yourself whether they read like you.
For the full list of tells and what to do about each one, read the words that instantly signal AI-generated text and delve, tapestry, and other AI vocabulary to avoid. If your score is coming mostly from em dashes, our why AI loves the em dash piece explains where that habit comes from, and the free humanizer prompt will clean up most of it in one pass.