[ad_1]
A student at Princeton University created an app that helps detect if a text is written by a human or using the artificial intelligence tool ChatGPT.
As Edward Tian, a senior computer science major, recently said: Tweet The algorithm behind his app, called GPTZero, can “quickly and efficiently detect whether an essay is written in ChatGPT or by a human.” Get the beta version of the app here. can.
ChatGPT has skyrocketed in popularity recently because it allows you to spit out coherent essays on almost any topic in just seconds. The technology has attracted investor interest, with The Wall Street Journal reporting that his parent company OpenAI could soon attract an investment worth $29 billion.
But the app also raised concerns that students were using the tool to cheat their homework, leading New York City to block access to ChatGPT on school devices and networks.
All it takes is a simple prompt like “Write a summary of the American Revolution.”
Tian now offers teachers everywhere another weapon against ChatGPT.
Usage
Tian’s app probes two variables in any given sentence under investigation. First, we look at the “confusedness” of the text, which measures randomness. Human-written text tends to be less predictable than bot-created work.
Tian told CBS MoneyWatch:
Tian explained that if the text looks very familiar to the GPTZero app and is less “puzzling”, it was likely machine-generated.
In addition to perplexing, his algorithm looks at “burstyness,” which measures variance or inconsistency within the text.
“There is a lot of variability in human-created articles,” he said.
“Man deserves to know the truth”
According to Tian, more than 20,000 users have demoed GPTZero’s beta, and teachers around the world have thanked him for making their jobs easier.
“They said it worked, which confirmed the suspicions they had.
Students, on the other hand, are less appreciative of Tian’s efforts.
While AI takes hold, Tian believes safeguards are essential “to ensure that these new technologies are adopted responsibly.”
‘I’m not against ChatGPT at all, but humans deserve to know the truth about what AI is and what it is human-generated,’ said Tian.
[ad_2]
Source link
