Another Grading Crisis Hits College

Commentary

How has the arrival of artificial intelligence affected the college experience? It’s worth asking because we now have myriad tools available that can generate college papers in seconds. A dissertation is a click away. Even whole books with deep citations can be constructed in an hour.

It’s not as easy to tell real from fake as it was a year ago. The language engines are getting more sophisticated. They can easily be customized to appear more real than they are. And although using them in place of your own work might run contrary to institutional rules, there is nothing illegal about capturing this content as your own.

A new working paper from the University of California–Berkeley has documented at least one effect of this. The number of A grades given in classes affected by AI has gone up by one-third in just two years. It seems like everyone is smart, writing perfect papers, passing tests, and getting all the problems correct.

“Courses with more AI exposed tasks, such as writing and coding, saw substantial grade increases after ChatGPT’s release,” Igor Chirikov wrote. “The share of A grades rose by 13 percentage points, or about 30 percent relative to the 2022 baseline. These increases were larger where homework carried greater weight, consistent with AI substituting for student work rather than broad learning gains from AI.”

Students who eschew the use of AI are in the trap. They see their colleagues getting away with it, getting good grades, and saving their nights and weekends for other pursuits. You join them or get eaten in a system that does not seem to reward honest work. This breeds a wild cynicism.

When I was in college, a major advantage of joining the fraternity was to have access to rooms filled with tests that had been pillaged from professors via their teaching assistants. I always found the practice disreputable.

Indeed, I worked with my professors to ferret out plagiarism among my fellow students, mainly because I found the practice morally outrageous. I would take a stack of papers to the library and trace footnotes, often finding that one-third of the papers (in an honors program) had sections that were entirely fake.

The problem from my point of view was not the violation of copyright; it was the pretense of knowing what one does not know. That’s a form of fraud. I admit to experiencing great joy in policing the classroom for this sort of thing.

Today, matters have changed entirely. Every student has access to all the answers from his phone or laptop. They see nothing wrong with taking recourse to this, all in the interest of getting the grade and ultimately getting that piece of paper on graduation day. The whole system is rather dark, in my view, but all these tools defeat the entire point.

What’s more, the students know that the professors themselves use AI to prepare the tests, so why shouldn’t they use AI to prepare the answers? That’s how they think about it and justify it. So we have AI grading AI, just as speechwriters (I’ve noticed) are using AI to generate press briefings. It’s truly difficult to tell what is and is not human anymore!

Of course, this problem in college has led to yet another round of complaints that grading is impossible in the university setting. Oral examination is the best and ultimate way to test a student’s knowledge, but that is not practical for faculty. There isn’t enough time in the day.

I suspect that there is something else going on here. Many college professors these days face enormous pressure from the deans and administrators to give high grades even when they are not merited. Kids are borrowing six figures and living off their parents to buy a product, which is not wisdom, but a degree. Any professor who is grading down is certainly going to run afoul of the administration and not get promoted.

As a result, over decades there has developed this university culture of wink and nudge, as if everyone knows that the system builds in plenty of toleration of cheating but no one wants to talk about it because there is too much at stake. After all, keeping warm bodies around the campus and maintaining access to the money flows is how the whole institution survives.

I’m sorry to report the above, but I’ve seen more than enough evidence of the truth of this.

How would grading take place if the actual priority were on learning? This is not a mystery. Yes, there is the oral examination route, which lacks practicality. So let’s take a look at how private credentialing takes place in professional societies. Here we have extremely high standards that are maintained because the credibility of the certification depends on it.

Here I am speaking of the following: chartered financial analyst, certified public accountant, certified management accountant, certified internal auditor, certified financial planner, chartered alternative investment analyst, financial risk manager, certified treasury professional, certified fraud examiner, chartered global management accountant, fellow of the Society of Actuaries, associate of the Casualty Actuarial Society, chartered enterprise risk analyst, and so on.

There are many others besides in computers and tech. Every serious profession has a credentialing system in place, something about which college students are generally unaware. They get their degrees and discover that the barriers to success in any field are much higher than they thought. The college degree alone is not considered to be a demonstration of skill or worthy of trust.

These credentials are extremely difficult to acquire. They force massive study, and students often fail examinations on their first try. Preparing for them takes enormous discipline and study every day for months to a year. The pages that must be mastered can be in the tens of thousands.

But once the person has this credential, he is set for life. In this way, these credentials are far more valuable than the doctorate. Indeed, I would personally delight in putting the Harvard, Yale, and Princeton financial and economic faculties through just one of these days. I seriously doubt that there would be any passing grades.

So how do professional societies do it? They put all responsibility on the student to be a self-learner as guided with tutorials, videos, and other aides. They discover whether the material has been learned with multiple-choice exams that are exceedingly difficult and randomized per test taker to make them impossible to game. The testing rooms forbid phones and electronics. The students sit in cubicles and everyone is monitored with a camera.

In other words, it is possible to stop the cheating epidemic. All major professional societies know how to do it. It’s how entire professions assess quality, from finance to accounting to actuaries to computer coders. It’s not a mystery. The problem has already been solved. There is a reason that universities in general—there are exceptions—are not interested in the solution. They don’t want to know because they need to keep their customers happy.

That’s a hard conclusion, but it is the one that makes the most sense to me.

Views expressed in this article are the opinions of the author and do not necessarily reflect the views of The Epoch Times.