That Time ChatGPT Wrote My Column

And other thoughts on why humans are still useful.

July 2023

Reading time min

That Time ChatGPT Wrote My Column

Illustration: Sarah Hanson

Prepare, we were told. This time it’s coming for your job.

Writers and editors are accustomed to reports of our obsolescence being greatly exaggerated. For years, we’ve been subject to gotcha stunts: Can you tell the difference between the article written by a human and the one written by AI? (Admittedly, I usually guess wrong. But in these set-ups, both articles tend to be workmanlike, and—here’s the catch—rely on a set of previously known, rather than newly gathered, facts.) 

We are also a curious lot, and we got to wondering: In the age of generative AI, which of our human characteristics aren’t slated to be supplanted? So we asked a group of Stanford faculty to opine on four areas that are central to our well-being: careers, health, relationships, and creativity. (Have you seen the one where DALL-E tries to draw hands?)

In our story, professor of law and of political science Daniel Ho observes that one of the tasks still firmly in the province of humans is analogical reasoning, or the ability to apply principles from a known situation to a new situation. This is why we’ll still need lawyers, but it’s also why we’ll still need science writers, who frequently rely on analogies to explain complex discoveries to lay readers. In the May issue of the magazine alone, heart sensors became postage stamps, polymers became meandering country roads, and the discernment of a geological epoch became the plot of an Oscar-winning film.

I have not yet met the AI creation that can follow its nose, calling new sources to report on a historical subject in greater depth than has ever been done before.

Assistant professor of computer science Diyi Yang points out that as AI writes more and more of our emails and texts, we’re all going to start sounding the same. Oh, indeed. As ChatGPT put it when I asked it to take a crack at this column, “Striking a balance between leveraging AI’s capabilities and fostering genuine human connections remains a challenge.” Its points were broad, its platitudes numerous, its transitions facile. 

Perhaps the chatbot’s most incisive observation was this: “Creativity, often regarded as a uniquely human trait, faces both disruption and augmentation by AI.” True enough, according to associate professor of music Ge Wang, who teaches a course in which students use AI tools to make new compositions. But he also observes that part of what makes art great is the suffering behind it, and the stories we tell that illuminate that suffering.

I have not yet met the AI creation that can follow its nose, calling new sources to report on a historical subject in greater depth than has ever been done before (see our cover story on Sam McDonald). Or that can think, Given the happenings in Major League Baseball, wouldn’t it be great if we talked to members of the 1962 frosh and JV teams about their experience with speedup rules, and then, when the reporting strikes out, pen a charming story about the effort. Or write a sentence like “It was a questionable media launch: a podcast with a professor lecturing on complex biology for hours while his dog snored,” that makes you want to drop everything and read a profile of professor-podcaster Andrew Huberman.

Which means I’ll be writing this column again in September. I think.

Kathy Zonana, ’93, JD ’96, is the editor of Stanford. Email her at

You May Also Like

© Stanford University. Stanford, California 94305.