Artificial Intelligence
Is AI Really a Threat to Higher Education?
Peak AI may be near. Nurturing thinking skills in college will remain essential.
Updated October 19, 2024 Reviewed by Abigail Fagan
When handheld calculators were invented, schools around the world didn't stop teaching basic math. Today, a decade or more of most peoples' lives remain occupied with learning math skills that have been easily and cheaply automated for more than half a century. Indeed, with nearly all cellphones featuring a calculator function, you could argue that there has never been a time in human history when basic math skills have been more superfluous. Yet no one seriously argues that basic math should not be taught.
The same is likely to be true for whatever cognitive tasks AI ends up being good for. Of course, in tech circles, a popular tongue-in-cheek definition has long been that AI is "anything that doesn't work yet." But one task it is undoubtedly good for is cheating on written assignments in college.
The thing is that using AI to cheat in college is entirely self-defeating. College students are by definition trying to get ahead of non-college graduates in terms of economically useful knowledge and cognitive skills. The main skill they gain is what is often called "critical thinking," or as I like to call it, "thinking." So it is college students themselves who stand to lose the most from cheating with AI, which aims to do the thinking for you.
More specifically, it is typical college students, moreso than the highest or lowest performing, who stand to lose the most. If you are considering using AI to help you with an assignment against an instructor's instructions, you are only hurting yourself. This sounds scoldy but I mean it to be very practical advice, and it applies in different ways to all students.
First consider the top-performing students. Some have predicted that, as more cognitive tasks are outsourced to AI, there will be two classes of people: those who program the AI and those who are subject to it. This seems like an extreme vision but there is good reason to believe that top-performing students stand to benefit from not needing AI to succeed, and this may help them shape future AI.
But it is not just future AI programmers who stand to benefit from honing their cognitive skills in higher ed. It is the whole spectrum of knowledge workers from lawyers and doctors to teachers, managers, marketers, artists, and architects that will still need to use their brains to be creative and think critically. Yes, some circumscribed tasks in these jobs will devolve to AI, just as financial accounting has been largely outsourced to the proto-AI of Excel spreadsheets. But all of these jobs will still require and increasingly rely on the thinking skills taught in college. If someone is inclined to cheat with AI in college, it doesn't bode well for their future success in these kinds of careers.
Lower-performing students also stand to benefit from eschewing illicit AI assistance. In correcting my students' assignments, I've noticed that I now take flawed writing more seriously than I used to. In the past, I might subtract points for ungrammatical writing, even if the student included the correct material. Although effective writing is another primary goal of college education, I may have been too quick to dismiss such responses. Now, irregular grammar is a good indication that the student did not use AI assistance since AI is, if nothing else, a near-flawless composer of grammatical sentences. These students, like all others, surely have easy access to AI, so I infer that they consciously chose not to cheat, and hope that they have done so out of an earnest desire to learn.
They are better thinkers for it, and they will have a leg up compared to those who use AI to cheat.
However, this is not to say that AI should have no role in the higher ed classroom. Instructors will need to think about what AI interactions could be beneficial for students. But maybe more importantly we will need to think about what student-AI interactions can be tolerated.
We have past examples to follow: things like spell checkers and grammar checkers, which have been around for decades. I'm sure most instructors would be glad to know that their students could compose an essay with excellent spelling and grammar without assistance. But few if any would begrudge (let alone penalize) students for utilizing spelling and grammar tools in college.
We may find that the effect of AI is neither a pandemic of cheating, nor a transformative enhancement of our minds. Instead, it will be tolerable aid for specific tasks.