Using AI to cheat has become too easy - The Boston Globe

Iris Liao is a recent graduate of Newton Country Day School who will attend Boston University in the fall.

During an English writing assignment, a student glances around the room to make sure that the teacher is distracted before quietly pulling out his phone under the desk. He enters a test question into ChatGPT, and within seconds, an entire essay is generated for him. He begins copying the essay word for word onto his paper as if it is his own. Will the teacher figure it out?

That’s a question I’ve asked myself after witnessing similar scenes in my high school. There is growing evidence that this is happening in high schools across America.

With the rise of artificial intelligence, cheating has become part of everyday classroom culture, something students seem less ashamed of doing. Will schools respond with policies that ensure that students who cheat with AI don’t benefit at the expense of honest peers?

It is vital that they do, and not just to protect honest students.

In the before-AI days, cheating involved things like sneaking answer sheets into a test or copying answers from a nearby classmate, making cheating on unexpected exams or in-class assignments more difficult. But the new style of cheating, which can be applied to every question and every testing method, is worse for the cheaters themselves because of the way it undermines their learning.

More than those old forms of cheating, AI outsources the thinking process, making traditional methods of measuring student thinking, like essay-writing or showing work, almost useless. Students are no longer just finding shortcuts on tests but rather replacing the skill development that the assignment was meant to teach.

This new kind of cheating is difficult to prevent precisely because it is so easy. It has become normalized, an everyday part of school life, no different than study periods. Teachers try to convey the importance of academic integrity. But their reminders become meaningless when they have few foolproof methods of knowing whether students are actually being academically honest.

Even elite colleges are struggling to prevent this kind of AI cheating. Recently, Princeton University announced that it would assign proctors to watch students during exams because too many students have been using AI to cheat.

When one of the oldest and most respected universities in America no longer trusts an Honor Code that lasted more than a century, it reveals how difficult it has become for schools to maintain academic integrity.

This issue becomes increasingly worrisome when students are pitted against each other in high-stakes situations, such as college admissions.

Imagine two students applying for the same college. One stays up late studying for tests, writing essays, and struggling through assignments. The other finds a way to make schoolwork easier by using copying and pasting prompts and cheating through tests.

When senior year rolls around, both students have similar grades and transcripts, making them appear like equals to college admissions officers. The admission officer can’t see who put in more work because they can’t know what goes on behind closed doors.

So then what happens? Colleges will inevitably treat both applications equally, punishing the honest student for the dishonest one’s behavior.

Students will learn that those who cut corners are rewarded with the same opportunities as those who choose not to cut corners. Hard work will come to seem superfluous.

Many schools are trying to take action. Some have started assigning more in-class writing on paper instead of on laptops, making it harder for students to look up answers during assessments. Others have prohibited phones, smart watches, and tablets that can easily be used to access AI in class.

The policy at my former high school, Newton Country Day School, is that AI should be used as a support tool for improving productivity or brainstorming, rather than replacing a student’s own work. Teachers are encouraged to help students learn the difference between what is an acceptable and unacceptable use of AI.

Since the decision of using AI is up to teachers, some of the boundaries surrounding AI can be confusing. The line between enhancing student learning and outsourcing their thinking is not always well defined. Teachers should play a role in helping students understand the value of the first and the dangers of the second.

I think that it is fine for students to use AI in moderation. I use it as a tool to help me study, whether it is explaining homework questions or making mock quizzes to help me practice for tests. When it is late at night and hard to get help from teachers or peers, AI sometimes becomes the best source available. AI can also be useful to check spelling, find sources for research, and as a resource for studying. In these situations, students can benefit from AI and still be able to think independently.

Schools should have clearer policies that make their stance on AI more consistent, while at the same time enforcing stricter rules on assessments. The challenge will be finding a balance of allowing students to use AI, while trying to ensure they are really thinking for themselves.