Created by Edward Tian, a Princeton senior, GPTZero is designed to assist educators in identifying writing created by ChatGPT
GPTZero is a tool that was designed to identify writing produced through ChatGPT ChatGPT, an AI writing software that was launched in November. It caused waves of shock through the education system due to its capability to create text that appears human when asked for prompts.
GPTZero was developed in the hands of Edward Tian, a senior student at Princeton University who majors in computer science, and minors in journalism. GPTZero has been made accessible to users for no cost for educators and other professionals as well as detecting work produced from ChatGPT more than 90% times, Tian says in Tech & Learning. It is among numerous innovative detectors that has been developed after the release of ChatGPT.
Tian discusses how he came up with GPTZero and how it functions as well as how educators can use it to stop cheating using ChatGPT in their classes.
What is GPTZero?
Tian was prompted to create GPTZero following the release of ChatGPT was released. He along with many others realized the potential this technology offered to combat cheating by students. “I believe this technology will be the norm in the near future. AI will be around for a while,” he says. “But at the same time, we have to build the safeguards so that these new technologies are adopted responsibly.”
Prior to the launch of ChatGPT Tian’s thesis been focusing on the detection of AI-generated languages and he was employed in the Princeton Natural Language Processing Lab. After winter break, Tian found himself with ample time and decided to code on his laptop in cafes to see whether he could develop an efficient ChatGPT detector. “I was like why don’t I just build this out and see if the world can use it.”
Everyone has become attracted by it. Tian has been interviewed on NPR as well as other national publications. Nearly 20,000 education professionals across the globe, and from K12 to higher education have subscribed to receive information regarding GPTZero.
What is the process by which GPTZero function?
GPTZero detects artificially generated text by measuring two characteristics of text, namely “perplexity” and “burstiness.”
“Perplexity is a measurement of randomness,” Tian states. “It’s an indication that reveals how unpredictably or an item is the language model. If a bit of text is incredibly random, chaotic or is not familiar to a model of language, If it’s a challenge to the model of language It’s likely to have a high degree of confusion and most likely created by humans.”
On the contrary it is the case that text that is familiar and likely to have been viewed by the AI model of language before won’t be confusing to it, and tends to be created by AI.
“Burstiness” refers to the complicated sentence structure. Humans tend to vary the length of their sentences as well as write it with “bursts,” while AI language models are more concordant. It is apparent when you make a graph that shows the variation in sentences. “For a human essay, it will vary all over the place. It will go up and down,” Tian states. “They’ll be sudden bursts and spikes, versus for a machine essay, it will be pretty boring. It will have a constant baseline.”
How Can Educators Use GPTZero?
The pilot version for free of GPTZero is accessible to all educators via the GPTZero website. “The current model has a false-positive rate of less than 2 percent,” Tian states.
But, he cautions teachers not to interpret the findings as proof that a student has made use of AI in order to cheat. “I do not want anyone to make final decision. This is something I figured out during my holidays,” he says of the tool.
It also has its limitations. It’s for instance, it’s not intended to recognize a mixture of AIor human-generated text. Teachers can sign up to be on a mailing list to receive announcements about the next edition of ChatGPT that will be able to highlight parts of text that may originate from AI. “That’s helpful because I don’t think anybody’s going to copy the entire essay off ChatGPT, but people might mix portions in,” the researcher adds.
Can GPTZero Keep Up With ChatGPT As The Technology Improves?
However, even as ChatGPT or various AI language models advance, Tian is confident that technology like GPTZero or other AI-detection programs will catch up. “Training the detection model is much less difficult than preparing one of these huge language models. It costs millions and millions of dollars in order to build these massive large model languages,” He says. That’s why ChatGPT was not made during winter break at cafes that offer free WiFi such as GPTZero was.
A journalism minor and avid reader of human written words, Tian is equally confident that the human element in writing will continue to be important in the near future.
“These language models are just ingesting gigantic portions of the internet and regurgitating patterns, and they’re not coming up with anything really original,” the researcher states. “So being able to write originally will remain an important skill.”
Read More :-
- What Is Blooket And How Does It Work? What’s New? 2025
- What is kahoot ? how to kahoot.it join a geme 2024 ?
- Https://aka.ms/remoteconnect – Minecraft
- How does the rice purity test work 2022
- What is Flippity and How Does It Work?
- Frax: How to Use It to Teach Fractions
- what is vudu? how to use vudu login
- How to watch Hulu in Canada? [January 2025]
