As an immigrant to the United States, who came here from India more than 30 years ago, it has been obvious to me for some time that people having access to guns everywhere made everyone suspicious and no one really safe.
Though it is certainly not life and death, a parallel confusion is evident in discerning what is real and what is fabricated from news and social media to homework that is completed with little or more than a little help from AI. As a parent and university professor, the confrontation of truth vs. fake is nearly constant.
I grew up in Mumbai/Bombay, India where guns were not plentiful except on Bollywood screens. Parents could send their children on play dates without worrying about whether the adults in the house had guns.
Here in the U.S., it is different. Plentiful daily gun violence belied the use of guns for protection. Guns made mass shootings –where more than four people are shot –entirely possible and probable. In the U.S., there were nearly 340 mass shootings from South Carolina to Minnesota to Mississippi.
But beyond the encounters between a police officer and an individual out in the streets, a doorbell ring is taken to be a home invasion and one can never be sure who has a gun and who does not. We are suspicious of each other and have to overcome more than shyness to build community.
In the age of AI, one can never be sure what is real and what is fake, what is disinformation and what is truth. To be sure, mortality is not at stake with the ubiquitous AI use, but the future is.
In schools and colleges, the rise of large language models (LLMs) is making essay writing a breeze with just a prompt. I have struggled with assessing how much a student has written is their own work. Now that the first semester of the year is half completed, it is critical to examine how to deal effectively with AI use before more knowledge learning is lost.
It is necessary to treat LLMs as new ways of accessing the digital world, replacing search engines, much like the calculators replaced the slide rules. A 2025 study of more than 12,000 high school and college students shows that 97% of the students use AI in their assignments. It should be noted that not all use of AI is cheating.
As in other domains, we are using AI to solve problems created by AI. It is possible to run the term papers through various plagiarism-detection software and decide to set the checking software at the level of a phrase or of a sentence, or at the level of an inarticulately paraphrased but correctly cited scientific paper. Then, assessing if say a plagiarism score of 10% (on some hastily defined scale) is sufficient to initiate punitive actions. There is little guidance from college administrators or the developers of such software packages.
Perhaps developers are waiting for the real-world training instructors inadvertently provide to improve their packages. Perhaps in less than a year, they will post data with confidence intervals based on the behavioral patterns of students and teachers.
To be fair, cheating has always been prevalent in schools and colleges. And, instead of bribing a classmate to write your essay, you need only write a prompt in ChatGPT. As always, it is up to the instructor to decide how much thought, agony and learning went into a term paper.
But we lose something when we begin each encounter with a stranger or read the first lines of an essay from a place of suspicion. Instructors wonder if the student did really write this, how much came from them and how much was an LLM, and more importantly, do the students understand what they have written.
There are some dire warnings of the use of AI. As recent research shows it can result in, “The erosion of trust, the hollowing-out of the student–teacher relationship, and the further commodification of learning threaten to precipitate a legitimation crisis for the academy itself.”
One solution may be to go old school on this one and use pen and paper in class. And also to really get to know the students IRL.
