Why Do College Programming Professors Require AI Now? A Parent’s Firsthand Experience
Last month, I was folding laundry at our kitchen table while my 20-year-old son Javi, a second-year computer science major, picked at leftover carnitas tacos between homework. He looked up from his laptop halfway through a stack of mismatched socks and said, “My data structures prof is making us use AI for every assignment this unit. Wild, right?” I immediately dropped a sock. My first thought was the same one I bet a lot of parents have: Isn’t that cheating?
I asked him to show me what he was talking about, and he pushed the laptop across the sticky table. The assignment header was clear, broken into four steps: Step 1: Write a prompt asking an AI tool to generate a working implementation of a merge sort algorithm for unsorted linked lists. Step 2: Run the code, identify at least three functional or efficiency errors the AI made. Step 3: Document each error, explain why it occurred, then rewrite the code to cut runtime by at least 20%. Step 4: Compare your final code to the AI’s original, write a 500-word reflection on what the AI got wrong and what you learned from fixing it. One hundred percent of the grade is based on steps two through four. I leaned back in my chair, surprised. I’d expected to see an assignment that let kids turn in AI-generated work for a grade, not this.
We sat and talked for another 20 minutes while he waited for his test runs to finish. I remembered when I took intro comp sci 25 years ago, I spent 12 hours trying to debug a simple program because I missed a single equals sign in an if statement. I was so frustrated I almost changed my major before I even got to the part where I had to actually understand what the program was doing. Javi said that’s exactly what his prof told the class on the first day of the unit. Most students who drop out of comp sci programs don’t leave because they can’t think algorithmically. They leave because they get stuck on trivial syntax errors before they ever get to practice the actual problem-solving that the job requires.
I was still a little skeptical, so I emailed the prof a few days later. I didn’t ask for any special favors for Javi, just wanted to understand the reasoning behind the new requirement. I got a reply back the next day that was straightforward and easy to follow. He said that since AI tools became widely available, he’s tracked performance between two groups of students: one that did assignments the old way, writing all code from scratch, and one that used AI for baseline code, then spent their time debugging and optimizing. The AI group scored 15% higher on average on the cumulative final exam—which is still closed-book, no AI allowed. He also said that every entry-level programming job he’s seen in the last two years expects candidates to know how to use AI productively. If his students graduate without that skill, they’re at a real disadvantage compared to graduates from other programs that teach it.
It’s not all perfect, of course. Javi told me three students in his class just turned in unedited AI code with a generic explanation, and all of them got failing grades. The prof doesn’t care if the final code works. He cares that you understand what you changed and why. Misusing it the same way people misused CliffsNotes back in the day still gets you in trouble, just like it always did.
What I’ve learned from this experience, that I’ve passed along to other parents of college students I know, is pretty simple. First, instead of asking “Is that allowed?” the second your kid mentions AI for an assignment, ask “Can you show me what the assignment asks you to do?” That one small question keeps you from jumping to conclusions that aren’t true. Most of the time, it’s not cheating—it’s just a different way of teaching the hard stuff. Second, if your kid is studying programming or any tech field, ask them to walk you through how they use AI. Javi showed me how he tweaks his prompts to get better baseline code, what common errors he looks for first, and that explanation process actually helped him catch one error he’d missed himself. It’s a win-win: I stop worrying, he gets to practice talking through his work out loud, which reinforces what he’s learning. Third, just check if the class still assesses basic skills without AI. Javi’s class still has midterm and final exams with no AI allowed, so I don’t worry anymore that he’ll never be able to write code from scratch when he needs to. If a class requires AI for everything with no closed-book assessments, that’s a red flag, but that hasn’t been Javi’s experience so far.
I still have small questions sometimes. Last week Javi told me his next unit will require AI to help outline project architecture, and I wondered again if we’re giving up too much of the old way of learning. But then he showed me the mental health app he’s building for his senior project, how he uses AI to handle the repetitive boilerplate code so he can spend time working on the accessibility features he cares about for teens who can’t afford therapy. It’s too early to tell what this shift will mean for a whole generation of programmers, but what I see right now is that my kid is spending more time problem-solving and less time staring at a screen crying over a missing semicolon. That doesn’t feel like a loss to me. It feels like a change, one we’re still getting used to, day by day.
Please indicate: Thinking In Educating » Why Do College Programming Professors Require AI Now