My students cheated... a lot

Yes - it's similar to the situation in high school where there are kids who want to learn and kids who disrupt the class because they're lazy/not motivated/have issues/want attention.

Automated teaching and testing with social sharing automate that dynamic. There's less attention seeking, but much more passive aggressive subversion.

But that's not the scariest thing. The scary thing is that if it's a STEM field these students go on to get jobs in which they have no competence. This is truly catastrophic if you want software that works and buildings that don't collapse.

Worse - the skill they've learned best is gaming the system and hiding their incompetence.

It's a double failure - of culture as well as knowledge.

Kudos to the prof in the story for handling it so well. Most profs won't.

The underlying issue is that there's been far too little research into the social consequences of automating all kinds of interactions.

The 70s utopian ideal of "Give everyone a computer to empower them" turned out to be ridiculously naive. What happened instead is that various dysfunctional economic and cultural patterns were automated and enhanced.

Culture as a whole has no defences against this because hardly anyone has realised that it's a problem inherent within the culture-amplifying effects of automation, and not an unfortunate byproduct that just sort of happens sometimes - and who knows why?

My position on this has been pretty controversial when I've shared it before, but I still think it's correct:

Measuring knowledge at scale is futile, harmful, and pointless. The fact that a lot of society has been arranged around the fiction that this is a feasible endeavor does not mean it has borne out in practice, and prioritizing assessment in this way has been gradually hollowing out most forms of pedagogy of their value while building an ever-expanding series of increasingly meaningless hoops for people to jump through to get what they actually need. We have deemed it necessary to create assessments to prop up the idea that education can be easily measured and should gate meaningful life outcomes for most people. Most if not all "cheating" behavior is either just a rational, strategic response to this situation, or a disconnect between how people actually solve problems (e.g. often collaborative and laser-focused on the part of the problem that drives the outcome, in this case the assessment) and some weird cultist notion of what it means for an individual to do it "correctly".

Effective pedagogy will never scale unless we get some really AGI-like technologies (I loved The Diamond Age as a kid, but A Young Lady's Illustrated Primer is from the perspective of extant tech a total pie-in-the-sky fantasy, illustrative of how meaningful teaching requires individualized approaches), and we see time and time again that teacher-to-student ratios as well as particularly good specific teachers are overwhelmingly the drivers of even the stupid metrics we are optimizing for

In short, this whole system is broken because its fundamental premise is flawed

Group work is often set up specifically to force this kind of "cheating". My course used it as a way to disguise the fact that students weren't learning programming. They tested everyone at the very start of the course to locate people who already were self taught, then assigned each group one self taught person who ended up doing all the work. The rest ride for free, everyone passes and the scam continues.

I've tasted university education very little ever since. Academia is just flooded with intellectual fraud of all kinds. The students sense the lack of seriousness and accountability and realize that no matter what is said, that'll get away with it because the profs don't really care and in fact depend on cheating to survive.

Not necessarily. During Covid my school (in France) shifted to online quizzes. Students would start by screenshoting the questions, which would then get automatically uploaded to a shared Dropbox folder. Students would then run a program that would cluster the screenshots and outputs the names of students who had the same questions and then they can collaborate directly. This is brilliant of course and solves all of the communication issues the students in this post faced but hearing about after it the fact (I was an international student and didn't know anyone in my class at the time) made me sick and salty. Most of those students are now workingnin high paying engineering jobs lol I am too but yeah, realistically, cheating works.

The reason why doctors have to jump through so many hoops is that the stakes are higher for failure.

While there are times where a piece of code failure or poorly worded instructions can cause injury to others, those are exceptions to the rule. Generally speaking- the cost of failure for writing and software is lower than it is medicine, and it makes sense not to gatekeep these industries behind theory, and rather just let results speak for themselves.