The AI writing debate is eerily similar to the one that rocked ancient Greece 2,700 years ago

Ancient Greece had a word for something that helps and harms at once: pharmakon. Plato used it to frame writing itself, and the argument sounds almost unnervingly current. Writing could preserve memory, or it could weaken it by letting people lean on marks instead of their own minds. That same split now hangs over generative AI: useful with judgment, hollow without it.

The article draws a straight line from that old debate to modern desks where speed matters. Circe’s warning to Odysseus — choose the lesser loss, not total ruin — becomes a useful image for newsrooms, law firms, and research teams deciding how much human effort to keep in the loop. The trade-offs are already visible. The Associated Press automated corporate earnings stories in 2014 and expanded coverage from roughly 300 companies per quarter to 4,400. Bloomberg began publishing headlines with machine learning in 2015 and, by 2019, was using automated tech to help with roughly a third of its content, according to The New York Times.

That bargain comes with a cost. Stanley Druckenmiller’s Wall Street Journal op-ed on the bond market was run through Pangram this August and came back as 100% AI-generated; he confirmed it, and the paper did not add disclosure. The point isn’t that the machine always writes badly. It’s that the line between assistance and authorship keeps getting blurrier, and the blur is the product.

The piece then reaches for Greek drama to make the point sting. Aristotle objected to the deus ex machina when a god arrived from outside the story to solve what the characters could not. Drew Lichtenberg, the Shakespeare Theatre Company’s artistic producer, flips that image: AI is less the machine than the god coming out of it, stepping into a role once held by people. That’s the real unease here. Not that AI is new, but that it keeps finding old ways to make human judgment look optional.

And the oldest warning still lands: a tool that extends thought can also displace it. The deadline crowd will call that efficiency. The rest of us should probably call it a loss, even when the spreadsheet smiles.