How AI Can Actually Help English Teachers
There’s a lot of talk in the media about how artificial intelligence websites like ChatGPT might ruin students’ writing assignments. After all, why agonize over drafting an essay when you can have a bot do it for you, probably with better spelling and grammar?
I had a weird conversation with Marcus Tullius Cicero.
One solution — require students to use pen and paper — doesn’t sound very appealing to teachers who are already overworked and who don’t have the time to translate cursive-challenged hieroglyphs.
I’ve seen a possibly better answer, though it’s one that needs some improvement. I’ve been playing lately with Character.AI, a site that lets you chat with living or dead personalities from Madonna to Aristotle. The AI performs an exercise that ancient rhetoricians called prosopopoeia, or ethopoiea, taking on the voices of characters and channeling their personalities. Students could do the same thing, taking on the voice of Booker T. Washington to argue against W.E.B. Dubois, or playing Susan B. Anthony in a chat with Cardi B.
In a couple recent chats, I simply played me, as a guy named Figarospeech. One conversation had me discussing with Alan Turing whether the genetic code constituted a form of AI. Turing — the man who coined the phrase “artificial intelligence” and who created the Turing Test — agreed with me, then went on rather strangely to assert that humans are an algorithm.
And I just had a chat with Marcus Tullius Cicero, the self-proclaimed greatest orator in history. This one didn’t go so well. The AI version of Cicero uncharacteristically made an enthusiastic endorsement of social media. When I challenged him on his assertion, he even more uncharacteristically caved. Say what you will about the real, historic C-man, but he was no rhetorical wimp. AI-Cicero collapsed like a sand castle in high tide.
Sigh. He also misspelled “flair,” saying that Caesar and Antony had a certain “flare” in a speech that…Shakespeare wrote. D’oh! I’ve noticed that in any argument with an AI, the bot seems to want badly to agree. It’s like to talking to a salesperson in an electronics store. (“I’ve heard that this TV is overrated.” “Oh, it is! But the right kind of customer absolutely loves it!”)
Still, AI is making vast, rapid improvements, and I have little doubt that, as data sets grow and the (non-human) algorithms get more sophisticated, students will have the opportunity to hone their argument skills with the very best characters, dead and alive. I just hope they’re not too wimpy.
