If you use Google Docs, pretty soon, there will be a new feature, Help Me Write, that automatically asks you if you want to use artificial intelligence to help as you compose your document.
“Once you request it, it's default, so it's there every single time you open a Google Doc,” said Jane Rosenzweig, the director of the Harvard College Writing Center, on GBH's Morning Edition. “You see this little magic wand and it says ‘Help Me Write.’”
Rosenzweig writes about what AI means for writing and teaching and recently penned an op-ed in the L.A. Times about the trouble with using writing assistants from places like Google and Microsoft.
She said the current suite of AI writing assistants is just the beginning.
“This is apparently going to be rolled out in every kind of writing software that you may use,” she said. “And my question is, are we going to still be doing our own thinking?”
Writing, Rosenzweig said, is not just the act of putting words on paper. It’s the process of thinking.
“What worries me as someone who has spent my entire career helping people learn to write is that when we open up a document to start writing, what we're really doing is beginning a thinking process,” she said. “So I use the example in my article: I was trying to figure out, what do I think of this new feature in Google Docs? I knew that it was bothering me. I wasn't exactly sure what was bothering me, and it was in the process of writing that op-ed that I really figured out what the concerns were.”
Google’s writing assistant comes preloaded with some suggested prompts: A thank you letter after an interview, or a social media caption for Employee Appreciation Day.
“I worry a little bit that we sound like we're having a great time here and isn't this fun? But that shouldn't detract from the real concerns that I think we all ought to have about outsourcing our thinking to a machine,” Rosenzweig said.
Rosenzweig said she can imagine “all sorts of horrible scenarios where we outsource all of our thinking.” But she also said she does not believe those scenarios are imminent.
“I think what I'm really reacting to here is that there's a thread of discussion about ChatGPT and generative AI that assumes inevitability, right? Get on the train. It's coming. It's here,” she said. “It might be a moment where we say, Do we want to use these tools? Do we understand exactly what's happening when we're using them? … We have to decide what these tools are useful for rather than letting the technology determine the future.”
To test it out, we asked it to write an introduction to a newscast for a journalist named Jeremy Siegel.
The program came back with a radio script: “Good evening. I'm Jeremy Siegel. And here are the top stories tonight.” It followed up with a quick rundown of headlines about the war in Ukraine, unemployment numbers and the Federal Reserve expecting to raise interest rates.
It ended with “We'll have more on these stories and more after this break.”
