The Paradox of Originality: How AI is Changing the Way We Write

Over the past year, I’ve had a sneaking suspicion that everything written, from articles to emails to LinkedIn posts, has grown exponentially longer. So, it was validating to read Victoria Turk's new piece in The Atlantic, "The Great Language Flattening," confirming this trend.

The research she cites is fascinating: when people were exposed to AI-written content, their own writing became 166% longer on average. As AI makes verbose content effortless to produce, will brevity become the true mark of thoughtful, human communication?

But there is an interesting paradox here. AI tools learned from human writing. The LLMs didn’t invent the em dash or the “not x, but y construction,” that is rampant in AI-generated text. It’s just parroting back what is already all over the Internet.

To fight back against writing like an AI attempting to write like a human (what one researcher quoted in the article calls an "uncanny valley" feel with "nothing behind the eyes"), some humans will, "deliberately lean into [their] idiosyncrasies" to signal authenticity and to create writing that's "the real deal," notes language evolution expert Simon Kirby in the piece.

But should good writing be defined by how successfully it avoids algorithmic patterns? I am not so sure. True distinctiveness isn't found in deliberate eccentricity, but in having something genuinely original to say. The originality lies in the human perspective guiding the process, not necessarily in the tools used to express it.

Perhaps in an era where anyone can generate endless paragraphs, the most revolutionary act is knowing exactly when to stop.