Recent developments in the world of Artificial Intelligence (AI) have caused a seemingly exponential uptick in interest surrounding online chatbots like ChatGPT, otherwise known as “Large Language Models” for their ability to replicate human written language, and generally very convincingly so. Language models are getting better quickly, too – with the release of GPT-4, which costs $20 USD/month, you can experiment with an even smarter bot. But with such a quick rise to fame, how should we slow down and think about potential consequences in the field of college life and education and  in the world at-large?

As a disclaimer, I did not use a chatbot to assist me in writing any part of this article; I do not trust it. As human writers, we all have our own styles of writing, and as college students, we are changing and perfecting them consistently, with or without realizing it. Yet, I cannot say I have not used a language model of some kind to help me. In typing this, every time a blue squiggle appears under a word prompting me to change it, like when I typed “getting better quick” in this article and changed it to “getting better quickly,” a language model is behind the scenes – no matter how careful I may want to be in writing this article without the help of a novel language processor, a language model has still inadvertently influenced the way I write. The reach of chatbots is already great, and will only continue to grow at an increasing rate.

Overuse of models like these will affect, maybe even slow the development of, our writing styles, language, and monographic idiolects. AI has already integrated with our human brains in substantial ways, from personalizing Instagram ads influencing the things we buy to recommending news articles and shaping the way we think or even vote, as has been the fear in recent elections around the world. Should we allow AI to influence our usage of human languages themselves?

One of the first places to ask this question is the classroom, where similar questions have often been asked before: is AI not akin to an abacus, or to a calculator in math class? In a recent article for Wenhui Daily, a Chinese paper, NYU Shanghai Chancellor Tong Shijun made a similar point. Yet if chatbots in this sense have the potential not only to be a writer’s crutch, but also something that artificially influences or even hinders the development of their writing style, is the problem posed more serious than the benefits to be gained?

I believe that these are valid questions, and certainly important ones to ask, especially as the power and ability of AI and large language models grows. As the influence of AI continues to climb, we have the duty in the classroom to think about its potential consequences. This does not mean ignoring it entirely – rather, we should entertain a philosophy of bringing the world to the classroom, as is a larger stated goal of our university, which requires taking a more liberal approach to matters such as these. A good education should attune students to what is happening outside of the four walls of campus, and train them to address the issues outside of university life.

Yet some conservatism in the classroom, I believe, is not entirely undue. As liberal as NYU may be, my most recent Social Science midterm was oral, Oxford-style, and I personally appreciate the moment when a professor asks the class to close their laptops to have an old-fashioned discussion or debate without the use of the internet. I’ve personally begun taking notes and even writing some essays in longhand in order to reduce distractions, and to force myself to slow down and think without being plugged into a computer. 

The same must be done, to some extent, in relation to language processing. In the same way word processors and the internet are great tools for education, AI should be used as helpful aids inside and outside of the classroom, but not totally relied on. Further developments in the technology will force humans to find the middle ground, and students and members of institutions of higher learning should continue to be leaders in that debate.