By Ava Bialow

ChatGPT is a thoroughly competent AI chatbot created by the company OpenAI. The “GPT” in its title stands for Generative Pre-training Transformer. As this title implies, ChatGPT can transform commands of any length into generated text, using previous requests and answers to guide new responses and perfectly emulate a human exchange. The chatbot is incredibly versatile in its abilities as it possesses endless amounts of data. Some of its skills include, but are not limited to: composing music, writing essays, coding and debugging computer programs, simulating chat rooms, playing games, and mimicking human conversation. OpenAI’s ChatGPT is revolutionary in the AI world due to its consistently accurate human-like responses. While the benefits of this remarkable creation are already being applied around the world, it does unfortunately, produce some conflicts, especially within school systems. 

As our world’s technology evolves and the presence of AI increases, new questions and concerns arise within school systems regarding academic integrity. ChatGPT can write essays in seconds on any topic, and a student can receive such work with the click of a button. As previously stated, chatbots like ChatGPT are so technologically advanced that their writing is almost entirely indistinguishable from a human’s. Therefore, teachers are challenged with the task of determining whether a student’s work is their true writing or rather an essay generated by AI. Accordingly, plagiarism policies in most schools are being reworked, edited, and altered to include chatbots as a form of plagiarism. 

However, a 22-year-old college student, Edward Tian, created a program that is able to successfully detect whether or not a piece of text was written by AI. Tian is a computer science major at Princeton University who created GPTZero this past winter. Within a week of the app’s launch, GPTZero had over 30,000 users. The app uses two main signs to indicate whether something is AI or not: perplexity and sentence-structure variation. If the bot is perplexed by the text, it is more likely to be written by a human and have exceeding complexity. On the other hand, if the writing is familiar to the bot, it is less complex and therefore much more likely to be written by AI. Additionally, humans vary their sentence structure in writing much more frequently than bots do, so a lack of fluctuation in sentence length is another indicator that the piece was written by AI. While this app isn’t perfect and makes errors every now and then, its accuracy is improving over time with alterations to its code as well as the new knowledge and experience the bot obtains. 

Ultimately, while AI chatbots open doors to new creative opportunities that can make regular tasks much easier, it’s important to set boundaries on how much society should depend on them in daily life. Additionally, it’s exceptionally interesting to see the ways that people use new AI to combat issues that other bots have created, proving that our society is technologically advancing by the minute in order to resolve conflict.