With the artificial intelligence making its way into our day to day lives, it is beginning to claw its way into academia. Breaking news has wreaked havoc throughout university halls when word got out that AI generated applications, such as ChatGPT, can write full-length essays for students. This new technology has made teachers and professors worried about the academic ramifications that this could cause students and educators alike. ChatGPT, which has experienced a meteoric rise in fame in the past few months, has become known for its ability to generate human-like conversations with people. Now it has the capacity to create essays and documents while posing itself as human.
The 2020s have already seen astronomical changes in the world of computers and technology. With VR headsets now changing the way we experience gaming and entertainment to Kiosks replacing cashiers at your local McDonalds, it is widespread knowledge in this point in time that technology has deeply assimilated itself in our daily lives. But with the admen of ChatGPT and its grand capacity to create words, sentences, and paragraphs out of thin air, it scary to think of the time of world we are entering. And it could cause big problems…
But with big problems come big solutions. And that is something that Edward Tian, a 22-year-old senior at Princeton University, is seeking to find. According to NPR.org, Tian has been working on an app, GPTZero, which will be able to autodetect any hint of AI intervention used any essay or document. Tian has been noticing the news buzzing around his Princeton campus and wanted to see what the hype was all about. (npr.org) “His motivation to create the bot was to fight what he sees as an increase in AI plagiarism. Since the release of ChatGPT in late November, there have been reports of students using the breakthrough language model to pass off AI-written assignments as their own,” stated the website. Plagiarism, which is to be considered as one the “deadliest sins” that a student can commit, has become common due to the earth-shattering technology.
According to NPR, the “anti-plagiarism” application measures the level of plagiarism by only two metrics: “burstiness” and perplexity. Perplexity measure how complex or diverse the written word is. If GPTZero detects high perplexity in a document, the more likely it is written by a human being. The converse would be true as well. The other indicator would be “burstiness”, which measures the number of long sentences compared to short sentences in any given essay. Human beings tend to write with high “burstiness” because we tend to write an essay with a combination of long and short sentences. If GPTZero detects low “burstiness”, then it means that the essay has more “uniform look” to it.
Mr. Tian has shown several demonstrations of the effectiveness of his application and it has worked for the most part. During one of his demonstrations, “Tian compared the app's analysis of a story in The New Yorker and a LinkedIn post written by ChatGPT. It successfully distinguished writing by a human versus AI.” This shows a bit of proof of the good that it is coming our way, technologically speaking. If there was any doubt that the AI bots were coming to destroy the world as we know, young professionals such as Tian, show us that we can indeed beat fire with fire.
Companies have been working around the clock to minimize AI-generated plagiarism. Some companies have thought of placing in an “invincible watermark” on AI-generated documents to help schools find the origin of the content.
Another intriguing part about Chat GPT is that it show signs of “hiccups” and mistakes that it creates for its users. An example of this is explained through the concept of “hallucinations.” According to ArtificialIntelligence-News.com, the technology engages in literary error when it starts to go off the rails on a topic that it was supposed to talk about. According to
Artificalintelligence-News, “The model will give you exquisitely written–and wrong–text about the record for walking across the English Channel on foot, or will write a compelling essay about why mayonnaise is a racist condiment, if properly prompted.” Clearly the technology is not all-too-perfect and does perform slip ups.
But the challenge will be when the technology gets developed to the point where the number of slip ups begin to decrease. According to the website, Chat GPT tends to have an error rate of roughly 20%. But what will be the answer once it begins to lower down to 15%, 10%, or %5?
That will be a bridge that will have to be crossed once we get there.
There is no doubt that the future of technology will bring about new questions surrounding academic integrity. Will this bring the end of the college education system? Or essay writing? Will this force colleges to bring in new forms of surveillance to the forefront?
Share This Post On
Leave a comment
You need to login to leave a comment. Log-in