With the advent of artificial intelligence, many applications and services have emerged that have a profound impact on our lives. Now that advances in AI technology are expanding exponentially, everyone is looking for shortcuts and answers to even fairly simple questions online to “save time”. Those related to academia, especially, have used online services like AI chatbots to find quick solutions, remove plagiarism, and more.
Now, with the advent of platforms like ChatGPT by OpenAI, which has been trained to produce long answers, many use it to absorb well-referenced and incredibly natural essays. Although OpenAI said it would flag ChatGPT’s output, people like Edward Tian are determined to solve this problem of AI plagiarism once and for all.
I spent the new year building GPTZero – an app that can quickly and efficiently detect whether an article is ChatGPT or human written
— Edward Tian (@edward_the6) January 3, 2023
Over the New Year holidays, the computer science student at Princeton developed an app called GPTZero. As the name suggests, this app can “quickly and efficiently” determine if your connection is creating ChatGPT. Tian showed off his creation in a series of tweets showing examples of GPTZero in action. One example showed that the New Yorker essay, “Frame of Reference” by John McPhee was composed by a human while a LinkedIn post was created by a bot.
GPTZero uses “puzzling” and “explosions” to determine text written by a bot. Bewilderment, in machine learning, refers to how surprised a language model is when it sees new data while the phenomenon that the use of a term once in a document is likely to happen again is called bursting. Human writing has more burst leading to more sentence diversity, and language models can predict text better when obfuscation is lower.
As Tian pointed out in a Substack newsletterOver 10,000 people have tested the publicly available GPTZero version in Streamlit To this day. He also mentioned that he updated the GPTZero model to reduce false positives. Furthermore, he is working on additional updates to GPTZero including new and scalable capabilities.
While Microsoft has plans to use ChatGPT’s capabilities to its advantage, the New York City (NYC) Department of Education has blocked access to ChatGPT on school devices and networks due to plagiarism concerns.
source: ribbed