How A.I. Can Help

“ChatGPT solves the blank page problem,” Cody Gough, a marketing professional outside of Chicago, told me. “The worst thing in the world is opening an empty document. ChatGPT helps you start.”

The limitations

As we’ve covered in this newsletter, A.I. chatbots are flawed. They often make mistakes — like the one that led to a $100 billion drop in Google’s cumulative stock market value when it appeared in an ad.

The workers I spoke to mostly use ChatGPT as a brainstorming tool and writing aid. They say the work of chatbots is sometimes inaccurate and often of lower quality than they could produce themselves, but that chatbots can still be useful.

Alexia Mandeville, a video game designer in Texas, uses ChatGPT to help her brainstorm character names, conceive ideas for trailers and produce news releases for her games. “I’m making something that doesn’t need to be factual,” Mandeville said. Because her work is fictional, she added, ChatGPT needs to be creative, not accurate.

The chatbot is still the tool, not the creator. It can copy writing styles, often replicating our weird internet behavior, as my colleague Cade Metz wrote. ChatGPT’s outputs are only as good as its inputs, so it struggles to reason, use logic, discern the truth and write imaginative work. It has tried and failed to write science fiction, for example. The human capacity for original thought is keeping white-collar professionals employed, even as A.I. poses more of a threat to them than earlier advancements did.

Freeing up more time

A.I. can’t do Snyder’s job of teaching music and gym classes. It can’t play the piano or the basketball game HORSE, and it can’t facilitate students’ social and emotional learning. But it gives Snyder more time for that work.

“Everyone is talking about how A.I. is going to replace us,” Snyder said. “I don’t agree with that. It’s going to free up more time at our jobs to do other, more productive things.”