@readmodelmag
MODEL

AI News

The em dash is in witness protection: writers scrub it to avoid sounding like AI

SummaryOne year after Sam Altman said ChatGPT would finally obey 'no em dashes' in custom instructions, human writers are the ones dodging the mark. Here is what the reporting actually shows.

· MODEL Editors · 1 min read

Some links may be affiliate links. Disclosure

Cartoon. Caption: Writers in 2026, scared of sounding like a chatbot. A nervous em dash character in a trench coat, fedora and sunglasses stands at a desk under a sign reading Punctuation Protection Program while a bored agent hands it a new ID. Bottom caption: You were used way too much by AI. From now on, you're two hyphens.
MODEL cartoon. Illustration by MODEL. Original. Image source.

The longest dash in the font has become the most suspicious one.

A year ago this week. On Nov. 14, 2025, OpenAI CEO Sam Altman posted on X: "Small-but-happy win: If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it's supposed to do!" (TechCrunch (opens in a new tab)). The fix was opt-in: it covers people who add the instruction, not ChatGPT's default output.

Now the humans are the ones hiding. Some examples from the reporting:

  • Novelists. WIRED reported in July that novelist Laura Brooke Robson took "a more winding path to avoid an em dash or a tripartite list" while writing her book, and that writers are converging on an "anti-AI" style built around avoiding obvious AI tells such as lists of three, "not x but y" phrasing and em dashes (WIRED (opens in a new tab)).
  • Editors. At ACES: The Society for Editing's virtual conference, there were two sessions on em dashes. Morgan Strunsky of Leff Communications described "flagxiety," the fear of having content flagged as AI-generated, and argued that cutting a well-placed em dash just to dodge suspicion makes writing worse (Poynter (opens in a new tab)).
  • Twist. Poynter also noted former New York Times copy desk director Merrill Perlman citing The Economist's report that ChatGPT now uses markedly fewer em dashes, and that only Claude uses them as much as people do.

This weekend the same worry is trending on X again, with posts about writers adding typos or skipping Oxford commas and em dashes so they don't get called "AI slop."

The MODEL take. Punctuation isn't proof. Models use em dashes because people have used them for centuries and the models learned from that writing. If a dash is the right mark, keep it. If you're checking whether something was AI-written, look at what it says: vague claims, sources that aren't there, and quotes nobody can trace are better signs than a long dash.

Get three tested tools every Friday

Free, weekly. Unsubscribe anytime.

Sources

  1. More Typos, Fewer Em Dashes: Writers Are Creating an Anti-AI 'Literary Counterculture'

    WIRED (Jul 29, 2026)

  2. AI learned how to write from humans. Now humans are trying not to sound like AI.

    Poynter (Sep 28, 2026)

  3. OpenAI says it's fixed ChatGPT's em dash problem

    TechCrunch (Nov 14, 2025)

FREE · EVERY WEEK

The top AI tools this week, tested.

Every week MODEL's editors pick the AI tools worth your time, try them hands-on, and tell you what's good, what's not, and who it's for. Free, in your inbox.

Free. Unsubscribe anytime. Our picks are editorial and can't be bought.