Why we still can't stop plagiarism in undergraduate computer science (2018)

We regularly catch cheaters in our classes; I do so every term and report all the cases I see up to the dean. In my experience these have not resulted in catastrophic declines in teaching evaluations; the few unhappy with getting caught cheating are drowned out by the 80-90% who don’t cheat.

It is actually critically important that institutions take cheating seriously. Rampant, well-known cheating can tank the reputation of a program or institution, or at the very least serve to cheapen the value of the degree.

Remove the incentive to cheat, and save yourself the time trying to catch it (and punish it, despite an uncooperative administration).

Plagiarism is now rampant - but for structural reasons. The university seemingly decided they want the money of foreign students, and started deciding that degrees from certain foreign universities counted as prerequisites for CS Masters programs.

A lot of these foreign students somehow have little programming skill despite undergraduate degrees in CS.

The cheating is obvious. She gives students 0's with no push back from administration if she thinks they are cheating. Some amazing stories:

- Students who turn in assignments with someone else's name of them

- Students who admit to cheating and do it again and again, never becoming more sophisticated at it.

- Students who get caught cheating 5 times and ask if there is any way they can pass the course.

- Students who despite being told there is no way to pass the course, request a call or in person meeting to discuss how they might be able to pass the course (she has never taken them up on this - one wonders what their pitch would be)

Cheaters who put a little effort in are undetectable. Or at least she suspects, but can prove nothing and grades the (not obviously plagiarized) code on its own merits. The general feeling is that barring these foreign students, cheating is there but not too widespread.

She does, indeed, have a bad rating on RateMyProfessor and does not care. Her administration still thinks highly of her because of her other work (such as running accreditation programs).

If you are a professor YMMV. This is how it is for my wife.

Some of them simply don’t generalize, like creating new homework assignments from scratch each semester.

As much as I'm loathe to suggest using AI in CS, that's actually something an LLM might be able to help with --- generating tons of pseudorandom variations on a theme.

I think making assignments completely for practice, and having 100% of the grade be the final exam, where there is zero computer use and pure brain use, outputting on pen-and-paper, will work. Anyone who tries to cheat will only cheat themselves and have the results of that exam reflect their actual understanding.

I plagiarized quite a bit in school. I'm not proud of it. Desperation and poor role models can create all sorts of negative outcomes, though. I was taught how to survive, not how to live ethically.

You can try to filter the plagiarists, sure. But uh, I'm not sure if it will work. The plagiarists are in league with each other.

My friends who teach used the same level of filtering in the past: They suspected a lot of students from copying from each other, but they only took action on the cases where it was so undeniably obvious that it would be an open and shut case.

In the era of LLMs, that degree of precision is completely gone. However, the other signal is amplified: They have more students than ever getting 100s on homework and completely failing any exams. In the past it seems that many of the plagiarizers at least learned something, however minimal, in the process of copying homework and then trying to turn it into something that looked less suspicious. The LLM users are so brazen that they just submit the prompt, copy the output, and call it a day.

When I studied applied physics eons ago we had tutorials in term time and a substantial essay to complete in the long vacation but these were between the student and the tutor and had no effect on the final class of degree; that was determined by a series of final exams (open note), the report of the final year laboratory project and its oral defence. My project report was more than a hundred typewritten pages with numerous diagrams and I had to defend it to my project supervisor and the head of department. I don't think a plagiarist could have done it.

Of course this is expensive, all the academics involved were permanent employees of the university, there were no graduate students doing slave labour tutoring.

In my opinion the plague of plagiarism is a direct consequence of attempts to get education on the cheap and of valuing the diploma more than the education itself.

I'm pretty sure that student incentives come from having a degree that will instantly put you above 90% of non-degree-holder applicants after graduation, especially in this market.

Faculty can't compete with this, no matter how much they want education to be the goal; it's not, for most grads, and teachers should blame companies for that rather than students.

Plagiarism can be countered in comp sci settings by having in-person coding and bugfixing exams, which would also prepare students for real-world work much better than the 'turn in a big (music player app written in java) project at the end of the semester' model that programming classes in my comp sci program used.

The answer here is surely a monitored environment. An online editor/tester/compiler that logs not only the resulting code, but the way they got there. Anyone who pastes in 99% and just renamed the variables is trivially discovered. This probably works for far more than computer science.

It's not just a stick to beat kids with. You can work back to actually see where they're falling down, do seminars where you can open your environment to allow TAs and others to work on a single codebase collaboratively. This must exist already, no?

Why we still can't stop plagiarism in undergraduate computer science - https://news.ycombinator.com/item?id=16651099 - March 2018 (118 comments)

For those who went to schools with strong honor codes, would you advocate that to others?

For example, did you find an environment of trust and respect, like you've not seen elsewhere?

For another example, do you have strong post-graduation alumni network, where someone in the alumni directory isn't just a possible foot in the door for sales ("hey, we went to the same school, can you grant me the courtesy of a call"), but that you can assume they are likely honest and have integrity?

Did it improve the quality of your education?

Were there downsides?

This is why homework are NOW being done in-class: no takehomes.

In short, weekly exams, on-paper, in-person, in-class, on teacher's test form paper.

No phone. Calculator permitted. Some allow open-book.

I too welcome back the 1980s.

Mimograph, anyone? * sniff * sniff *

As long as it's clearly communicated in the syllabus, should be fine. If identical code submissions are so common then everyone should be doing the same quantity of work on average and it shouldn't be an issue if you automatically get assigned bonus problems.

Early on I've made the mistake of sharing my solutions with people I knew. Unfortunately they kept sharing the solution too, and so on.

Two times I was pulled into a profs office after the relevant lecture, to be questioned about it.

After it became clear that I was the author, and what happened, nothing ever came of it (for me).

Ironically both times the copiers supposedly failed to remove the git repo that was part of the handoff, so it was primarily about verifying I was the original author.

Lesson learned: "invisible" watermarks work, because people are generally lazy (also don't share graded work, just offer to help)

Is the point of the homework to have students just do some work or is it a tool to encourage them to more deeply engage with the material? If it is the former, then what is wrong with using LLMs? If it is the latter, then students should be judged based on their understanding, which they have to demonstrate in a scenario where deception is very hard or near impossible.

The real problem is students not wanting to learn. And what is the point of educating people who do not want to learn?

Wait, why don’t I just ask the LLM from the get go?

The ships are all gone from the harbor. Luckily for justice, there won’t be any jobs anyway for those degrees. Something about the universe is just doing quite a rug pull on so many things, fascinating times.

Why are adults at a university doing graded homework? They're responsible for their own learning, and if they want to cheat, well good luck on your exam. Our exercises in math and CS always were a few years old and kind of cycled through, so just copying old solutions was trivial, but what for, you're only hurting yourself. People who didn't learn just bombed out at the end. A university education is voluntary, people are there by their own choice.

Honestly this seems almost entirely like a byproduct of the privatized education system. University as children's daycare where everyone needs to pass and be tutored because there isn't an incentive to just filter people out.

The commons is the value of a university degree. No, not the education! The diploma.

There will always be some fraction of students that genuinely learn and earn their diplomas. Industry rewards them with higher salaries because the diploma demonstrates that they’re educated and self-motived intelligent people — grade A employees in other words.

So of course, some students will cheat and gain a diploma through copying others. This is inevitable, like the drug trade or prostitution.

If the percentage is low enough, nobody cares.

Okay, but… what if it’s a little higher? Those students might be highly profitable foreign students just looking to get a piece of paper so that they can get a better life back home.

What about a bit higher? Now the university is making bank, the foreign students are funding a new physics lab and a new pool! Awesome.

What if it’s 90%?

Now… it’s too late to rock the boat. The emperors nakedness must not be revealed! If industry catches on that almost all of the diploma-wielding graduates are actually C- instead of A+ the whole thing will implode.

That’s why universities are “helpless” against cheating! They won’t fix it, because if they do, the music stops.

We’re nearly there now. I walked through my old university and saw maybe 5% locals. I hardly heard a single conversation in English. This place is now graduating students in bulk that are functionally illiterate (in English), let alone the subject matter.

The song hasn’t stopped yet, but AI is fiddling with the buttons on the stereo.

There are no chairs.

You study to learn. Assignments are to help you learn and evaluate yourself. If you skip them in whatever manner only you suffer. Noone else. When you finish your studies and go to work noone will care about the grade you got, especially if you can't code your way out of a wet paper bag.

No one is going to stop the student and ask them why they cheated even if they were practically being asked to eat a tomato soup with a fork. It's easier to hide the problem under the rug and scapegoat.

    #include <stdio.h>

During the exam professor asked me why I called arguments of main argv and argc and if they can be called something else, I did know the answer, but I didn't know exactly what continue does because I wasn't using it in my code.