An app developed by a Princeton University student helps determine if a text is written by a human or by the artificial intelligence tool ChatGPT.
Edward Tian, a senior cs major, stated in a tweet that his algorithm, GPTZero, can “quickly, efficiently detect whether an essay or article or any text is written by ChatGPT or human.” You can download the beta version of this app here.
ChatGPT is gaining popularity for its ability to generate coherent essays on any topic in seconds. Investors are interested in the technology, according to Wall Street Journal. OpenAI parent company could soon attract investment valued at $29 billion.
However, the app has raised concerns that students might cheat on homework. New York City blocked ChatGPT access on school devices and networks.
It only requires a prompt such as “Write a summary about the American Revolution.” Thanks to Tian, teachers worldwide now have another tool against ChatGPT.
Table Of Contents 👉
How This ChatGPT Plagiarism Checker And Text Detector Works?
Tian’s app analyzes any piece of writing it finds. It examines two variables. It first examines a text’s “perplexity,” which measures its randomness. Human-written texts are more unpredictable than bot-produced.
Tian said that if a text is extremely random and unfamiliar, and the app is confused, it is more likely to be human-generated. Tian said that if a text is familiar to the GPTZero App and has low “perplexity,” it could be machine-generated.
In addition to perplexity and inconsistency within texts, his algorithm measures “burstiness” or variance. He said, “There’s a lot of variation in human-generated articles.”
Humans have the right to know the truth
Tian stated that more than 20,000 people have tried the beta version GPTZero and that teachers around the globe have contacted him to express gratitude for making their lives easier.
Tian stated, “They say it works, and that has confirmed their suspicions.” On the other hand, students have not been as appreciative of Tian’s efforts.
AI is here to stay, but Tian believes safeguards are necessary “so these new technologies can be adopted responsibly.”
ChatGPT is not something I oppose. Tian stated that humans have the right to know the truth about AI and human-generated AI.