Replacing em dashes by hand is tedious and easy to miss. Here is how to do it automatically and correctly.
The fastest way: a cleaner with a dash setting
Paste your text into a tool that lets you choose what an em dash becomes, and pick comma. A good cleaner will:
- swap every em dash (
—) for a comma in one pass - also catch the en dash (
–), which find-and-replace usually leaves behind - keep number ranges like
2010-2020readable as hyphens instead of turning them into commas - remove the invisible characters that often come along with AI text
textscrubr does this: turn on Normalize punctuation, choose the comma option, and review the list of replacements. Commas are only used between words, number ranges become hyphens, and nothing that could change meaning is applied until you approve it.
Find and replace (one document at a time)
If you would rather use your editor:
- Copy an em dash character from your text.
- Open find-and-replace.
- Paste the em dash into the search field.
- Replace with
,(comma plus a space).
Two things to watch:
- This does nothing about en dashes, so search for
–separately. - Replacing a dash that was joining a number range will read oddly, so review ranges by hand.
Should it always be a comma?
A comma is the most natural replacement when the em dash is setting off a parenthetical aside, which is how chatbots use it most. But check each case:
- Setting off an aside: a pair of commas reads best.
- Joining two complete thoughts: a period or semicolon is cleaner than a comma.
- Trailing emphasis at the end of a sentence: often just delete it.
If you want one consistent rule applied to a whole document fast, the comma is the safest default, and an automatic cleaner applies it everywhere in seconds while leaving your lists and code untouched.