Educators concerned that the viral popularity of OpenAI’s ChatGPT will lead to waves of generic-sounding, mostly AI-written essays, could be reason to relax. Princeton student Edward Tian devoted part of his vacation to writing GPTZero – an application that can identify text written by artificial intelligence.
Tian posted a few proof-of-concept videos on Jan. 2 demonstrating GPTZero’s capabilities. First, it established that a human being a New Yorker article; then it correctly identified ChatGPT as the author of a Facebook post.
here’s a demo with @nandoodles‘s Linkedin post that used ChatGPT to successfully respond to Danish programmer David Hansson’s opinion pic.twitter.com/5szgLIQdeN
— Edward Tian (@edward_the6) January 3, 2023
Business Insider has more:
GPTZero scores text for its “perplexity and burstiness” – referring to how complicated it is and how randomly it is written.
The app was so popular that it crashed “due to unexpected web traffic,” and is currently displaying a beta signup page. GPTZero is still available to use on Tian’s Streamlit page, after the website’s hosts stepped in to increase capacity.
Tian’s motivation for creating GPTZero was academic in nature, about that he cried “AI plagiarism.” Tian tweeted that he thought it unlikely “high school teachers would want students to use ChatGPT to write their history essays.”
The creators of ChatGPT at OpenAI have their own concerns about how their product is being used. As the Guardian Last week, a researcher recently reported in a lecture at a Texas university that they “want it to be much harder to take a GPT output and pass it off as if it came from a human.”
According to the GuardianOpenAI is currently working on a feature to “statistically watermark” ChatGPT output so that machine readers can discover buried patterns in the AI’s text selections.