Classroom Conundrum: Does the Em Dash Signal AI-Written Work?

To em dash or not to em dash, that’s a writer’s dilemma these days. Ever since the punctuation mark became the “ChatGPT hyphen,” internet sleuths have begun calling out AI-written work based on this symbol alone. 

The only problem? Text with an em dash doesn’t mean it’s AI-written. It just means the writer understands sentence structure and punctuation.

Today, we’re taking a deep dive into this topic to explore how to use an em dash, how it got a bad reputation from ChatGPT, and what you can do to actually spot AI-written work or plagiarized content in your classroom—without relying on em dashes alone.

• What does an em dash really do?
• Why does AI love the em dash?
• Tips to spot AI plagiarism in the classroom—beyond the em dash
• How to prevent AI cheating and plagiarism before it starts

[What does an em dash really do?](id-what)

An em dash is like the Swiss Army knife of punctuation marks. It can replace commas, parentheses, colons, or semicolons to vary writing and sentence structure. Some of its most common uses include:

• Setting off parenthetical information for emphasis.
• Setting off appositives that contain commas to prevent confusion.
• Bringing attention to a list that comes before an independent clause.
• Signaling interruptions or changes in thought in creative writing.
• Introducing a summary statement after a word or phrase.

Because the em dash doesn’t appear on your computer keyboard (but is accessible with shortcuts!), people may overlook using it in everyday writing or texting. These dashes don’t get as much attention in casual writing as hyphens, en dashes (which you use for ranges, like 4–5 hours), and other punctuation marks.

[Why does AI love the em dash?](id-why)

Our bot friends like ChatGPT, Perplexity, and Gemini love a good em dash. But why? Aside from not having to remember the keyboard shortcut to use one, there are some other explanations as to why this punctuation mark is an AI favorite:

• Formality: Many AI models are trained on formal writing—like academic journal articles or research papers—to help them learn grammar rules and “proper” writing. You’re more likely to see an em dash in professional writing than you are in a social media post.
• Variety: AI models are trained to avoid repetition when possible. They may add em dashes to emphasize content, change the rhythm of a sentence, or indicate a tone shift.
• Verbosity: Since AI models are trained to be grammatically correct and polished, the content output is typically more academic and less conversational. This verbose style may use em dashes, bigger words, and longer sentences at the expense of clarity and readability.
• Data mirroring: Em dashes exist in all genres and styles of writing. Since AI pulls from massive datasets and source data, many models have learned that em dashes are just part of the “natural flow” of writing.

Em dashes alone aren’t a reliable sign that a text is AI-generated. AI mimics human writing. ChatGPT and other AI tools may even tell you this if you ask. Good writers use em dashes, and AI is trained to copy them.

[Tips to spot AI plagiarism in the classroom—beyond the em dash](id-tips)

If you can’t trust the em dash as your indicator for spotting AI-generated or plagiarized work, what can you look for? Here are some tips to help you spot AI-generated content from your students:

• Trust your instincts: You know what punctuation knowledge on-grade-level students should have. If they’re using em dashes or any other grammar that’s too advanced or wouldn’t indicate gradual improvement, you’ll recognize it.
• Look for deeper understanding: Does the writer truly understand what they’re sharing? AI-generated content may provide a lot of facts, but not a deeper analysis of the topics or ideas.
• Question perfect work: Polished writing is a great goal, but inhumanly perfect texts with no typos, misused words, or any other mistakes may hint at AI generation.
• Check for personal insights: Writing that lacks personal details or text connections, or relies only on vague and generic examples, may signal AI content.
• Watch for hallucinations: AI-generated content may include made-up facts, sources, or other information.
• Spot repetitive patterns: AI-generated content often uses similar sentence structures and predictable word choices. If the written work deviates from your students’ unique writing style, it could be AI.
• Pick up on AI phrasing: While the em dash isn’t a reliable clue to spot AI-generated writing, overly formal or unnatural words and phrases like “foster” or “weaving a tapestry of collaboration” should set off AI alarm bells.

[How to prevent AI cheating and plagiarism before it starts](id-how)

It’s great to have a plan for how to spot and address AI plagiarism in your classroom, but wouldn’t it be better to prevent it before it starts? Here are some tips to help you address AI-cheating and use it as a motivator for students to do their best work—without relying on AI to do it for them:

Start a dialogue

Addressing AI cheating and plagiarism at the beginning of the year or at the beginning of a writing assignment is beneficial. It tells students that you’re aware of and will be looking for AI-generated submissions. But this also opens the door for a dialogue that allows students to ask questions about what counts as cheating or what uses of AI are allowed in your classroom.

To start the conversation and keep it going, try:

• Sharing your school or district’s cheating policy with students and reviewing acceptable and unacceptable AI use.
• Be clear about the expectation for every assignment you give, including guidance on what students can or can’t use AI for.
• Promote an integrity culture where you value original thought, learning, and individual progress over scores or letter grades.

Rethink assessments

Students may be most tempted to use AI to cheat or plagiarize when under pressure. This can happen when they have an extensive assignment or one that counts for a large portion of their grade. Try these tips to revamp your assessments and assignments to make them less cheating-friendly:

• Add visual or interactive questions: These elements are harder for AI tools to copy or interpret.
• Use open-ended questions: Scenario-based or critical thinking questions force students to demonstrate individual understanding rather than relying on regurgitating facts.
• Time your assessments: Using reasonable time limits makes it challenging for students to cheat with AI because they won’t have a chance to type every question in and wait for a response.
• Try a lockdown browser: Tech tools, like Formative’s Respondus Lockdown Browser, can prevent students from leaving the assessment window, limiting their ability to cheat.
• Schedule frequent assessments: Using smaller, regular formative assessments can reduce stress and students’ temptation to cheat on larger exams.
• Personalize assignments: Tailor tasks to students’ experiences, interests, or background to help reduce the urge to cheat.
• Assign projects: Real-world, hands-on assignments encourage critical thinking and problem-solving. These assessments are harder to pair with AI assistance. 
• Try human-centered assessments: Classroom debates, mock trials, and other collaborative projects are other ways for students to show that they know, and rely on “doing,” which makes it harder to cheat with AI.
• Document the process: Ask students to upload their work step-by-step, submitting pictures or documents of their process that leads up to the final project.

Reframe the conversation

Though we’ve mentioned its potential drawbacks, AI isn’t the enemy in the classroom. It’s a helpful tool for self-monitoring, time-saving, and organization when used correctly. Rather than taking an anti-AI stance with students, reframe the conversation. Teach about the benefits of AI vs. the drawbacks, and when it’s an appropriate time to use these tools.

• Add AI literacy: Instead of banning students from using AI, teach AI literacy. You can add it to digital media literacy lessons you already teach.
• Model responsible use: Show students how to use AI responsibly and teach them explicitly how and when it’s appropriate to use AI assistance.
• Teach the basics: Teach traditional punctuation rules alongside your AI literacy lessons to specifically address concerns about the em dash.
• Encourage AI criticism: Ask students not to take AI results at face value. Teach about hallucinations, fact-checking, and ways to verify if AI is telling the truth.
• Acknowledge helpful AI tools: Stress the difference between AI replacing original thoughts and work vs. AI supporting self-guidance. For example, tools like Grammarly may rewrite content for you, but tools like Newsela Writing help students learn based on their own drafts and mistakes.

Use AI the right way in your classroom with Luna

Newsela and Formative’s AI-powered teaching assistant, Luna, is an AI pal you can trust. Build assessments, create graphic organizers, translate texts, level articles, and more quickly and easily.

Whenever you see the ✨or a “Luna” tag across our products, you’ll know you’re harnessing an AI tool powered by 10+ years of education and learning science expertise. Want to know more? Discover the latest features Luna powers to make your prep and classroom experiences even better.