Amid the celebratory lauds for new AI text generator ChatGPT, more than a few people have wondered what effect these types of systems will have on academia.
How easy will it be for students to effectively plagiarize their work via AI generation? Is there anything educators can do to head that off? The problem seems almost insurmountable.
Oh wait a second this one kid in Toronto figured it out in like three days:
Over the last couple years, [Princeton University senior] Edward [Tian] has been studying an AI system called GPT-3, a predecessor to ChatGPT that was less user-friendly and largely inaccessible to the general public because it was behind a paywall. As part of his studies this fall semester, Edward researched how to detect text written by the AI system while working at Princeton's Natural Language Processing Lab. ...
After the fall semester ended, Edward traveled home to Toronto for the holidays. He hung out with his family. He watched Netflix. But he couldn't shake thoughts about the monumental challenges confronting humanity due to rapidly advancing AI.
And then he had an idea. What if he applied what he had learned at school over the last couple years to help the public identify whether something has been written by a machine?
Just pause for a moment and consider that at a time when most young students would be comprehensively vegging out, sleeping until 1pm and binging Halo Infinite, this young man "couldn't shake thoughts" about one of the most unique and pressing problems facing society right now.
That says a lot about a guy!
Edward already had the know-how and even the software on his laptop to create such a program. Ironically, this software, called GitHub Co-Pilot, is powered by GPT-3. With its assistance, Edward was able to create a new app within three days. It's a testament to the power of this technology to make us more productive.
On January 2nd, Edward released his app. He named it GPTZero. It basically uses ChatGPT against itself, checking whether "there's zero involvement or a lot of involvement" of the AI system in creating a given text.
Just so we're clear: This is indeed a pressing problem. ChatGPT hasn't been out for two months yet and students are already trying to use it to do their work for them.
It's obviously needed. But perhaps most poignantly, in addition to being able to detect fraud and cheating in schoolwork, Edward believes this app will serve a more noble purpose to that end:
"We're losing that individuality if we stop teaching writing at schools," Edward says. "Human writing can be so beautiful, and there are aspects of it that computers should never co-opt. And it feels like that might be at risk if everybody is using ChatGPT to write."
