Is it AI or just good writing? The em dash debate explained
Are you someone who second-guesses using an em dash — or think you can spot artificial intelligence just by how a sentence reads? In the age of generative writing tools like ChatGPT, even something as minor as a dash is sparking significant debates online. Some say it’s a red flag for AI. Others say it’s just good writing.
Why the em dash is suddenly controversial
Online discussions have taken a sharp turn against the em dash. Users across Reddit, X and LinkedIn have said they can “spot” AI writing when someone uses an em dash.
One LinkedIn user writes, “If I see an em dash I’m calling you out,” with many other users posting similar conversation-starters.
Some believe the punctuation has become a stylistic giveaway — especially in writing produced by generative AI tools like ChatGPT. However, professional writers say the em dash has long been used to add flair, context and rhythm to sentences.
Journalists, in particular, rely on it to insert supplemental facts or to include another facet into a sentence that already contains a comma-separated list. Many creative writers use the em dash in place of parentheses or commas for dramatic effect.
AI suspicion vs. punctuation literacy
Not only is AI suspicion becoming a hot topic online, but schools and universities across the country are still learning how to handle its use by their students.
Aileen Gallagher, a journalism professor at Syracuse University, told Straight Arrow News they understand AI is becoming more prominent. While her university isn’t ignoring AI, teachers are trying to find ways to utilize AI to help students with their workflow and manage their time more effectively.
When it comes to spotting AI in work, she says it has nothing to do with punctuation.
“If I suspect AI,” Hallagher said, “it’s because you’re using vocabulary that I don’t think you understand, or you’re referencing concepts we haven’t talked about in class.”
Gallagher said it’s more about context, not style. Others, including Oregon State University instructor J.T. Bushnell, agreed.
A punctuation mark with personality
Bushnell believes the em dash’s reputation as an AI giveaway is unfair — possibly just a symptom of its growing popularity.
“I think part of the reason this trend has taken off is because the em dash is both a cool punctuation mark — it’s really versatile, it’s really elegant-looking — and it has a name that clearly describes what it is,” he said.
The em dash is roughly the width of the letter “m.” The en dash, slightly shorter, indicates ranges in numbers or time.
Did ChatGPT get trained on em dashes?
Some online are wondering if ChatGPT’s frequent use of em dashes is a “quirk in the training data.” In a February discussion post on OpenAI’s community forum, a user questioned whether the model had developed an over-reliance on the mark due to its dataset.
Gallagher notes we don’t know exactly how large language models like ChatGPT were trained. But she’s concerned about the monotony AI can create.
“If everybody is using the same tools to synthesize information, everything starts to sound the same,” Gallagher said. “We know audiences don’t like that.”
AI as a tool — not a crutch
Instead of replacing the writing process, some educators are exploring how AI can fit into a writer’s workflow.
Bushnell has researched how creative collaboration — even with AI — can improve writing.
“The way collaboration works is when you ricochet off someone else’s idea into a new direction,” he explained. “AI can be incorporated into the writing process in the exact same way.”
That approach, he says, helps with creativity — as long as the AI isn’t doing all the heavy lifting.
Still, not all uses of AI are creative or collaborative. In 2023, a Futurism report found that Sports Illustrated had published articles written by AI-generated authors without disclosing that information to their audience. The fallout was swift, and the head of the magazine was fired as a result of the controversy.
“If you’re using generative AI to write or create something, you have automatically violated The Society of Professional Journalists’ code of ethics,” Gallagher said. “Because you’re essentially fabricating or plagiarizing.”
Many newsrooms have internal guidelines against the use of AI-generated content without a disclaimer, if not banning its use for creating copy entirely.
AI is everywhere, but readers still prefer authenticity
Unbiased. Straight Facts.TM
More than half (52%) of U.S. adults use large language models like ChatGPT, which are powered by AI.
According to a May report from Exploding Topics, ChatGPT is now the eighth most visited website in the world — with 4.57 billion monthly visits. That puts it above Amazon and X in global traffic.
More usage doesn’t mean more trust. Bushnell said most people don’t enjoy reading AI-generated content.
“The best way to cope with AI, if you feel it sucking you in, is to think not about your experience as a writer, but the experience of the person on the other end,” he said.
“Almost everyone hates reading AI — because it feels like the person producing it didn’t care enough to put in the work themselves.”
Rethinking creativity in the age of AI
According to a Pew Research study, most American workers said they don’t use AI much or at all in their jobs. However, among AI users, 73% are under 50 and 51% have at least a bachelor’s degree.
As AI becomes more common in classrooms and workplaces, Gallagher urged students and professionals alike to think bigger.
“If generative AI is doing the work for you, then what are we all doing here?” she said.
That question, she adds, extends beyond journalism — and should be at the center of how we all approach our work.
Ella Rae Greene, Editor In Chief
