Why College Professors Require AI For Programming (And What It Actually Means For Your Student)
The kitchen table was still sticky with syrup from Saturday pancakes when my 19-year-old son, a sophomore computer science major, pushed his half-eaten pepperoni pizza aside and tapped his laptop screen toward me last month during spring break. “Did you see this? My data structures prof says we have to use AI for every assignment this semester. If I turn in all original code I wrote by myself without it, I can’t get higher than a C.” I was folding his clean hoodies, one still stained with coffee from a fall hackathon, and my first thought was the same as any parent’s: That can’t be right. Isn’t AI cheating?
I’ve taught intro technology classes at my local high school for 18 years, so I know a thing or two about how coding is taught, but this was completely new to me. Three years ago, I was proctoring end-of-year projects where we told kids any use of ChatGPT was an automatic fail. Two years ago, my own son got a failing grade on a lab because a TA flagged his code as AI-generated, even though he’d written every line himself at 2 a.m. in the library. Now the entire rulebook has been flipped.
I asked him to pull up the full syllabus to make sure he wasn’t misreading it. It wasn’t a typo. The professor wrote, “By the time you graduate, every entry-level programming job will expect you to use AI as a daily tool. If we don’t teach you how to use it well here, you’ll be behind before you even start applying.” That stopped me short. I’d always thought of AI as a shortcut, not a required job skill.
He walked me through his first assignment for the class to show me how it works. The goal was to build a custom sorting algorithm for a dataset of user health information. He started by sketching out the logic on his whiteboard propped against his bedroom wall, listing all the edge cases he needed to account for: duplicate entries, missing data, sorting by multiple categories at once. That part he did completely on his own, no AI involved. Then he used GitHub Copilot to write the 20 lines of boilerplate code that formats the input and sets up the output file. That’s the part he says he used to spend three hours on, just getting all the brackets and semicolons in the right place, and it never taught him anything about how sorting actually works. When he got stuck on an edge case where duplicate entries were throwing off his logic, he asked ChatGPT to explain three different approaches to fixing it, then picked the one that worked for his specific dataset and rewrote the core logic himself. Finally, he had to turn in a one-page breakdown that says exactly what AI did, what he changed, and why he made those choices. That breakdown is 40% of his grade, just like the working code.
I started thinking back to when I taught algebra in the 1990s, right when graphing calculators became a required tool for class. Parents were furious back then. They said kids wouldn’t learn how to graph lines by hand, that they’d be dependent on calculators and never understand the core math. Now we don’t even think twice about it. We know that doing the repetitive calculation by hand doesn’t teach you the core concept—understanding what the graph means, how to interpret it, how to apply it. This is the exact same shift with programming.
A lot of parents I talk to panic when they hear their kid’s professor requires AI for coding, and that panic makes sense. If you went to college before AI, you assume any help with work is cheating. The first concrete step you can take is don’t jump to conclusions based on a 30-second phone call or text from your kid. Ask to see the full assignment breakdown and grading rules. Most of the time, it’s not “use AI to do your homework for you.” It’s “learn to use AI as a tool, the same way you’ll use it in a job after graduation.” The second step is ask your kid to walk you through their process. When my son showed me his one-page breakdown, I could see immediately he was doing more problem solving than I ever did when I learned to code 20 years ago. I spent 80% of my time fixing syntax and boilerplate, and 20% actually thinking through the problem I was trying to solve. Now he spends 80% of his time on the problem, and 20% on the repetitive setup.
I talked to a mom at my high school’s PTA meeting last month who was considering pulling her daughter out of her university’s comp sci program over this exact requirement. She said she was paying $40,000 a year for her daughter to learn to code, not have AI do it for her. I told her what I learned from my son, and asked her to check if her daughter’s professor required documentation of how AI was used. She checked, and sure enough, the class had the same one-page log requirement my son’s class does. She stopped worrying after her daughter walked her through the first assignment.
It’s still weird, if I’m being honest. I grew up learning to write every line of code from scratch, and it’s hard to unlearn the idea that any help is cheating. My son still gets in arguments with his uncle, a senior developer who started coding in the 1980s, who says all this AI use is making kids lazy. We don’t know yet how this will turn out. In 10 years, we might find out that a whole generation of programmers can’t write basic syntax from scratch if AI goes offline, or we might find out they’re building better, more creative products because they don’t waste time on the boring stuff. So far, what I see is that my son is spending more time working on side projects that matter to him—right now he’s building a free app for the local food bank to track their inventory—because he doesn’t have to spend all his free time stuck on stupid syntax errors. It’s a big change, that’s for sure. But it’s not the cheating nightmare a lot of parents make it out to be. It’s just education catching up to how work actually gets done now.
Please indicate: Thinking In Educating » Why College Professors Require AI For Programming (And What It Actually Means For Your Student)