Those Em Dashes Everywhere: How AI is Quietly Changing the Way We Write (and the Copyright Law)

A year ago, I started noticing something odd in social media posts. More and more people—people who never used Em dashes before—were suddenly using them. Once I noticed it, I couldn't unsee it. And it wasn't just coincidence. It felt like watching AI's fingerprints spread across our writing in real-time.

It makes sense. AI-generated text often follows certain stylistic choices—Em dashes, structured lists, a particular rhythm of phrasing. People using AI for content generation (consciously or not) might start absorbing these patterns into their own writing. At first, it annoyed me. Not because I dislike Em dashes, but because I could see AI’s fingerprints on posts that were supposed to be personal. Was everyone outsourcing their voice to AI without realizing it?

What's weird though—I've started liking Em dashes. I like the way they create emphasis, how they keep thoughts connected without the abrupt feel of a period or the hesitation of parentheses. So now I want to use them. But if I do, does that mean AI changed my writing style? And, honestly, does it matter?

This leads to a bigger question: Where’s the line between using AI as a tool and letting it reshape how we express ourselves? The U.S. Copyright Office has been actively examining the legal implications of AI-generated content. Their latest AI Report - Part 2 (thanks, Doug Hohulin , for sharing the update) says AI-generated work can be copyrighted—but only if humans add enough creative input to make it genuinely their own.

In other words, if you just let AI write something and hit publish, you can't claim it as yours legally. But if you reshape it, refine it, and inject your own creativity, then it counts. It's a fascinating line in the sand, because it validates something we've all felt: creativity isn't about the final product alone—it's about the messy human process of making something yours.

Beyond copyright law, the usage of AI tools for content creation is shaping policies in academic and research communities. ArXiv, for example, one of the largest repositories for scientific papers, updated their policy to allow AI tools for editing and improving language but makes it clear that AI can’t be listed as an author or used to generate substantial portions of content. Similarly, the Authors Guild provides best practices for writers navigating AI-assisted creation.

Essentially, AI can help polish the writing, but the intellectual work must still come from humans. It’s another version of the same question: how much AI assistance is too much before the work stops being yours?

The spread of AI writing tools brings up another worry: are we all starting to sound alike? If AI favors certain patterns and phrasings, and we're all using these tools daily, what happens to the quirks and imperfections that make writing personal?

It goes beyond just Em dashes. I’ve noticed AI-generated text often relies on the same phrasing patterns—"in today's fast-paced world," "now more than ever," or "the key takeaway is..." It’s polished, efficient. Sometimes when I'm reading online, entire paragraphs start to feel like they came from the same source, just with different nouns plugged in. If AI is everywhere, are we moving toward a world where personal writing styles blur into one AI-optimized voice?

I've found a few strategies that help me stay authentic while still taking advantage of AI's capabilities. First, I use it mainly for research and to explore different angles of the topic—never for the final version of the writing piece. I've also started paying attention to my writing habits (the good ones and bad ones). Instead of letting AI shape my writing, I’ve learned to use it as a tool to improve my own style.

I suspect we're moving toward a future where AI becomes just another writing tool, like spellcheck or a thesaurus. But how we use it now, while it's still new, might set the tone for whether our writing stays diverse and personal or becomes a sea of perfectly optimized sameness.

So yeah, those Em dashes are growing on me. Though I still can't help wondering what other AI-driven writing habits I'll find myself adopting next year. Maybe by then we'll all be starting sentences with "Indeed," or ending every paragraph with a rhetorical question. At least we'll be in it together. ;-)

P.S. And just to close the loop—did AI help with this blog? Sort of. I used it to test some phrasing and check grammar, but the thoughts, structure, and flow are all mine. Which, I guess, is the whole point of authorship.