If you read an article on AI, you’ll see many comments asking why students “can’t do their work without using AI anymore.” Nostalgia-filled commentary about how the comment’s author had to do X, Y, or Z without AI pervades the comments, along with a sentiment that it was much better back in the day. Then, you might see a comment about how AI is everywhere and we cannot combat it. Then, someone will imply that AI detectors aren’t accurate and that there is no way to teach us our way out of this AI world.
First, we disagree that AI detectors aren’t accurate. AI Detector Pro updates its algorithms daily and regularly conducts market research to evaluate whether we accurately detect original content. Our technology team runs every piece published on this blog (original) through our detector for training.
Second, our team believes that “we can’t teach around AI” to be a failure-forward mentality. Professors have dealt with many other technological innovations. Why are they refusing to adjust to AI? Internet commenters call students lazy, but is it the other way around?
Third, many departments will lose their student population if they refuse to adjust to AI. In a previous blog, we cited that students are increasingly skeptical of the value of a college education. Meanwhile, entry-level jobs now require generative AI skills. For decades, colleges sold their value on getting ahead in the workplace. Academics are now hobbling their graduates. Did you know that at one point, students memorized logarithmic tables? Did MIT graduate less intelligent engineers once calculators and computers came along? Or did they become more efficient?
Here are our thoughts on how an undergraduate college can teach around AI with a few simple adjustments.
Start grading on a Curve
Grading on a curve means professors will only give a certain number of people in the class an A. It is also one of the most effective ways to teach around AI. The US is notorious for grade inflation, where most people in the class get a high grade. Since professors repeatedly state that AI decreases the quality of student writing and makes it mediocre, they should stick to their guns and lower the grades for mediocre papers. All they need to do is transparently declare a standard of “mediocrity.” They don’t even need an AI detector to do this. They should grade the papers themselves and make an educated decision about what is and isn’t mediocre. OR they should do the work and set the grading standards for the teaching assistants.
Make papers a smaller portion of the letter grade
Do you know that papers are not the most significant portion of the grade in many colleges worldwide? If academics believe that student papers will never be free of AI, they should reduce the weightage of papers in the final grade. Make it about the exam. Sheesh, this isn’t rocket science. Transparently declaring what separates a B paper from an A paper encourages motivated students to achieve an A in the way the professor expects.
Learn from other departments that have always allowed open-book exams, like law schools
We should include law schools in this discussion. Law schools allow in-person open-book computer exams. The school invests in software that effectively blocks the computer from the internet. If students don’t take the time to make an outline and understand how to craft an argument, they will probably get a terrible to mediocre grade. Law schools grade on a curve. A student’s performance on the written exam determines the final grade. Undergraduate education doesn’t need to be this drastic, but arguing that one can no longer teach a new generation of students is silly. Like law professors, undergraduate professors must take the time to develop innovative exams that test how students think.
Ultimately, professors must change how they teach and grade. They must understand that the skills required in yesterday’s economy won’t serve all students in today’s economy. It’s simply better to outline open and transparent policies for AI use and grading standards. This may cause tension during a transitional period, but students paying $75,000 yearly for education deserve professors who try to keep up with the times. Until then, get yourself an AI Detector Pro subscription for that professor who won’t change their mind about AI in their classroom.