To teach in the time of ChatGPT is to know pain

LLM use is the most demoralizing problem I’ve faced as a college instructor.

See full article...

 

[methodmadness00]

The irony of better technology making online/asynchronous learning and many kinds of remote work untenable is not lost on me.

 

[xoa]

There's evidence that this is not correct. The act of writing seems to enhance learning in and of itself.

I don't see anything you linked saying this is the case for final exams? Which is what I was clearly and specifically was talking about. And any theoretical advantages are irrelevant if it means I am physically incapable of doing it and thus fail? Setting aside other issues like reading someone's handwriting.

Yeah I absolutely do still handwrite my own scribbles to myself and draw stuff etc. But the only consumer of that is me. The only direction for handwriting in terms of communication with others has been firmly to get rid of it. One of the last remaining hold outs I can think of was/are doctor's scrips, where I remember studies show it's resulted in a lot of incorrect prescriptions over the decades that can have deadly results.

 

[WereCatf]

I’ve been teaching college Earth science courses as a part-time faculty member for a long time now, all while juggling other jobs. I started because it was enjoyable; no one gets into this line of work for the famously poor pay or complete lack of job security. Working with students is just one of those genuinely fulfilling experiences that is addictive enough that they ought to warn people about it.

But thanks to generative AI, it has become mostly miserable―at least in certain settings.

It probably won't help you feel any better, but I'll say it anyway: some of us do very much value teachers and all the time and effort you put into sharing your knowledge and skills. What you do still matters even in this age of slop, don't let the grifters and propagandists put you down! It can't be easy for you guys, but I hope you can find a way to power through.

 

[GlockenspielHero]

I work in academic technology at a college and I'm currently both doing focus groups with students on AI use as well as designing a series of summer workshops for faculty on AI in the classroom.

The most salient comment I've heard in the past year was from a frustrated student who works hard on papers only to see other students get a higher grade when they spend 10 minutes with ChatGPT because "ChatGPT gives the answers faculty want"

This reveals a serious disconnect between what faculty want (student learning) and how they assess that. We default to grading term papers because those were reasonable instruments to see student capabilities- we don't actually care about the term papers themselves. (We're more than happy to throw then in the trash after a long weekend if grading). We just want to know if a student actually understands what we've been teaching

This is not a new problem- students have always been able to bypass writing papers by getting other (paid) people to write them -LLMs just make that easier and cheaper. The answer isn't to give up or to add some ridiculous amount of anti cheating tech, it's to change how we assess learning. Does that mean in-class handwritten assignments in a blue book? Maybe, if that's appropriate. Oral exams? Sure. But get more creative- I've used serious role playing games where students reenact historical events like the trials of Galileo and interactive simulations to let them play around with how an industry changes over time. One of my guys runs our innovation lab (makerspace) and he's amazing at figuring out ways to get hands on activities mapped to learning goals.

Is reworking your entire syllabus to deal with AI going to be a pain? Yes. Is it going to take more time? Probably. But it's not impossible to figure out new ways to assess learning- and who knows, they might be a lot more interesting than grading yet another stack of mediocre term papers

 

[Cloudgazer]

On the one hand I worry that my job may be eliminated by LLMs.

On the other hand I worry that I will be doing my job till I die as the next generation of people who might have learnt how to do it will incorrectly assume they can always get an LLM to do their work for them.

I'm not sure which possibility worries me more.

 

[shoe]

People using AI aren’t doing work. They’re commissioning it, just like a rich patron commissions art. Of course they learn nothing. They did nothing.

I don't know. There's a skill and a certain learning that goes into recognizing that Leonardo is Leonardo. There were and I'm sure are also a lot of wealthy patrons who threw their money away garbage.

 

[NedKrist]

However, if everyone else is producing heavily tool-assisted work (even if they barely understand the output), which is slick, competent and looks professional, then there’s a possibility that honest, thorough output worked through from first-principles will no longer stand out, and may even seem lacklustre by comparison.

Our experience has been that the floor has been raised tremendously, but good students (or job candidates) using LLM assistance are standing out less--regressing to the mean, as long as humans are putting in the time to read submissions. Those that that stand out as unique (whether truly human-generated or just with a better prompt, who knows) get more attention, even if not technically as polished.

 

[Cloudgazer]

I don't know. There's a skill and a certain learning that goes into recognizing that Leonardo is Leonardo. There were and I'm sure are also a lot of wealthy patrons who threw their money away garbage.

There is a real debate over whether Charles Saatchi recognised Damien Hirst's artistic genius, or whether he used money and his skill as an ad man to essentially create the brand of Damien Hirst, artistic genius.

 

[Geebs]

You obviously haven't been keeping up with exam cheating technology. We've long since passed the age of notes written on arms or legs. There are whole kits you can buy that let you connect miniature wireless cameras and speakers to your phone, so someone can see the question on your exam and dictate an answer to you (no doubt using an LLM to generate text these days). The wireless ear pieces in some of these kits are so small, you need a magnet to retrieve them from your ear canal. You can bet that those taking the Korean civil service exam (and those having to invigilate it) are very much aware of these kits.

Errrr… just ban phones from the examination room?

 

[KonaKat]

The last online class that I took we were encouraged to use LLM's on some assignments as long as we documented everything. We could use a cheat sheet for the exams and they needed to be hand written. The professor also required that we take photos of our cheat sheets and upload them for review.

I came away from that class knowing a lot more than I thought I would learn. I am able to use the knowledge I came away with a few times a week. The class gave me a chance to critically think when using Python.

For online classes, what about doing hand written timed exams, and then upload photos of the work to the professor? Could that work?

 

[ZippyPeanut]

While it's probably true that AI is not good for education, A1 on the other hand will prove to be very beneficial: "A school system that's going to start making sure that first graders, or even pre-Ks, have A1 teaching in every year. That's a wonderful thing!... It wasn't all that long ago that it was, 'We're going to have internet in our schools!' Now let's see A1 and how can that be helpful.... Every school should have access to A.1." (Linda McMahon, US Secretary of Education)

 

[LG11]

I'm in higher education (social sciences) and very much agree with everything Scott wrote. There are no short cuts in thinking, reading, analyzing, processing, and interpreting the world around us. This is obviously lost on many students who can turn in a whole series of papers within an hour of light editing LLM output. At our university, we returned to pen and pencil exams, and to oral exams.

"Why not embrace LLMs?" you may ask. We do, to a very limited extent. It has utility. None of my colleagues are anti-tech. The problem is on the student-side: we are not teaching and testing them, we are just wading to shed loads of LLM-output to find that one insight that is genuinely produced by students learning something. The very fact that students themselves do not see the problem is problematic enough. And yes, I generalize to 'students' because we also have evidence that a vast majority of the students do not know how to do without. Humbug.

Thanks for the write-up, @ScottJohnson It made me feel slightly better

 

[Tactical Finesse]

Errrr… just ban phones from the examination room?

There are smart-glasses rental services for exams, now. Because LLM-enabled smart glasses with cameras and screens are expensive, so "entrepreneurs" are buying them and renting them out by the hour for students to cheat.

 

[Rector]

Errrr… just ban phones from the examination room?

Sure but you have to be able to find them. Banning electronics and also constructing exam rooms as Faraday cages might work.

 

[DCStone]

Errrr… just ban phones from the examination room?

Unenforceable, especially at scale. (Do you want to be responsible for collecting and returning 1k+ cell phones per exam?). The phone doesn't have to be visible at all for the system to work, just near enough. And never mind the diabetics who use their phone to check blood glucose readings and control their insulin pump...

At my place of academentia, we're not allowed to have students leave their bags at the front of room anymore - they're allowed to keep them under their seat. (If you're wondering why, imagine a single phone going off during an exam in a backpack in a pile of 1k+ backpacks...)

And before you suggest it no, blocking cell phone signals for all exam rooms is not practical (nor legal where I am).

 

[NedKrist]

Even in this essay where "instructors hate this" is your central thesis, it's hard to connect with you emotionally since the persuasive part is all about student learning. There isn't really an explicit argument that making instructors miserable is bad. I wish you had told us more clearly that we should care about teachers being miserable.

I'm willing to accept that there may be a day where we don't have to worry about this because dedicated teachers are no longer necessary, but I think we're still at the point where we should take seriously a factor that drives people away from a profession dedicated to preparing high quality humans in the next generation. Can't back that up with data, however.

 

[Marlor_AU]

Secure testing can be (and has been) attacked as "inequitable". It is "unfair" to students with anxiety disorders, with ADHD, from underprivileged backgrounds, cultural minorities, etc. etc.

This is a really difficult issue to solve, because when policies are put in place to address the genuine needs of disadvantaged students, those policies are often actually exploited by students who are most advantaged.

Here in NSW, we have some robust and generous policies to assist disadvantaged and disabled students. Scribes, stress breaks, additional writing time, and so on.

However, this assistance is disproportionately used by students in wealthy areas and by private schools. One elite, exclusive private school had 26.3 per cent of students claiming assistance, while in Sydney’s disadvantaged south west, only 2.5 per cent claimed similar provisions.

Wealthy, advantaged students have the means to clear the hurdles to be granted assistance. Genuinely disadvantaged students simply can’t manage to get the paperwork in order or secure appointments with sympathetic specialists to gain recommendations for assistance. It’s become a tool that actually exacerbates the gap between advantaged and disadvantaged students.

 

[zeromind]

I am a university lecturer struggling with the same issues though fortunately I do not have online courses. I have been studying and thinking about this problem fro the past 15 months and while I don't have a solution, I have a plan.

Educators cannot be passive observers of LLM technologies. They must fully engage with it to understand how this tool like others in history -writing, books, calculators, the Internet - can be applied to the task of education. In that spirit I use the tool to generate course material, critique my work, etc. I first learned how to use the tool effectively for my workflows and needs. I have been exploring the weaknesses and strengths of LLM technology and where it adds value and where it distracts or harms.

Then the educator needs to be involved with the LLM tools the student is using. Provide prompts or agents designed to guide the LLM to constructive help for the student. Techniques such as scaffolding elements of an assignment the student already has demonstrated or providing meta-cognition support with reverse prompting or guiding through a particular pattern to a class of problem. Keep the AI from doing the object of the lesson and instead direct it to support the student. Then followup with periodic in-person reviews of assignments referring to particulars the student answered. The same with test questions: ask questions forcing the student to answer from their submitted work. Make it easier for the student to use AI constructively than to use is harmfully. LLM's are like chainsaws: You can cut down a forest with one or create artwork, but the quickest thing you can do with it is cut off your leg. The difference is education about the tool including practice, guidance, and review by an experienced instructor in a suitable environment.

The goal is to seek alignment with LLM in course material. It took education over a decade to figure out how to use the Internet in education and many would say we still have much to learn there. It is frustrating and hard, but education has never been easy. The same tools that bring us frustration do offer some relief.

Time and peer review will reveal if it is a solution. For now I am relieved to at least have a plan.

 

[Dachshund]

I am going to hazard a guess that a majority, perhaps an ultra super majority, of Ars readers have embraced some form of life long learning.

The challenge is to somehow educate folks, which may well be a majority of the population, that the friction of learning is worthwhile and beneficial to everyone.

In my opinion, this is the greatest challenge humans face by an enormous margin.

 

[andyallcorn]

Great piece. Particularly like the pithy responses to common justifications.

It takes most of us a long time to realise that the end product of any assignment is not the point of schooling, but rather being shaped by the process of creating it. Of course, it’s exactly that shaping that enables you to eventually understand this concept.

Is it possible that in LLM-driven education (and workplace), you can bypass enough of the process that you never realise you don’t even know how to think?

LLMs can be a useful tool, but I see them from the position of someone who was already taught many of the lessons that I could imagine them depriving me of.

 

[TimeToTilt]

I think the issue with LLMs is that we have spent centuries telling people one of the biggest signs of academic and subject matter credibility is quality grammar and writing. And ai is very very good at putting out a decently high quality sounding paper, regardless of the truth of that paper. And humans just aren't capable of dealing with that in our current framework.

Oh also, even when some of us did cheat back in the day, rewording Wikipedia or spark notes or whatever still took effort and importantly still exposed you to the subject matter. You actually did have to read the Wikipedia entries etc and know them well enough to reword, ai gives you a paper on seconds that you can just turn in and never even glance at. That is important I think

 

[cp1919]

Just go back to grading fully based on supervised written exams.

I used to teach an introductory Astronomy class - I would have two groups of students: the Star Trek Geeks who would sit in the front row and ask me questions about black holes and the Big Bang, and the other half sprinkled thought the room who were just there to get a required science credit.
It's hard to balance that separation when creating tests, and in grading fairly. You have to make it easy to get a C for those not prioritizing and hard to get an A for really motivated students.
Also, a large percentage were math-phobic or not good at essays, so you can't just give one kind of question - weekly quizzes, multiple choice, short answer, essay, extra credit, office hours
And I myself hate high-stakes examination, so I'm not going to make everything based on one last test. ugh.
I was very forgiving, especially if I saw any effort

As a non-major course, this was not important pedagogy, but the hope is to broaden the base and provide something interesting.
It was all in-class and turn in paper back then: to do it now would be very different. I'm not sure how I would restructure things, but I think the only solution is to try and incorporate the LLMs into the answering process. (which has it's own cost if there are still some not using it!)

I always started out telling them they were adults and if all they wanted was a grade that was OK, but the point was to actually learn something, and tried to keep them interested.

ultimately, you can't help everyone. My dad used to say: you can't teach anyone anything. they can only learn.

 

[Zeppos]

Teacher here. When I was 14 years old, I was surprised when a teacher told us we had it harder than they. My father often told me about how he had to study a literature book out of his head. Exam? You came in, teacher read a few sentences and you had to continue. Our teacher claimed we had much more diversion than they had. This was 1995. No internet, no smartphone. We had TV. Roughly 5 channels, some kid programs on a few of those for half an hour a day.
These days, kids constantly have to resist. Resist to binge watch something on netflix. Resist to engage in a hot topic interaction on whatsapp, resist to play free games. For schoolwork? They have to resist to look for the solution online, resist to ask the solution from eachother, resist to letting an AI do all the work. It is too demanding. They need help.
That is not a teacher's job. That is called parenting.
Of course, that is of little help when you teach an online course and students have to give in their homework for grading. Back to school kids...

 

[moriad]

Education has value, but apparently students aren't seeing it.

And why is that? Because universities have made it clear that they are profit seeking entities designed to squeeze every last cent out of students and their families while education is treated as an accidental side-effect of the process. The widespread belief in grade inflation means that anything less than an 'A' is a failure. And when they finally do end up with a job (probably not in their field), they'll be told by a cynical middle-manager that almost nothing they learned in university has any relevance to the work they're supposed to do.

There's a perception out there that university is largely a scam*. If that's really what you believe, why wouldn't you cheat?

*I got my degree sometime during the late Iron Age and students (and hiring managers) were already pretty cynical. Still, to me, cheating seemed an obviously self-defeating short-cut. But then I wasn't looking at a permanent lifelong 6-figure debt either.

 

[xoa]

Goodhart's law: When a measure becomes a target, it ceases to be a good measure.

Like so many other "laws" (ie, "useful aphorisms", including "move fast & break things" or "obscurity is not security") on the internet Goodhart's should be taken as a starting point for reflection and critical thinking in context. Not merely trotted out as an end in and of itself, or else it risks it risks becoming a bromide. Goodhart's law would be better written as "when a proxy measure becomes a target, it ceases to be a good measure", which both gets to the point but also helps consider that sometimes proxies are the best we've got. Sometimes one just needs to measure better.

We have culturally allowed grades to take on very outsized importance in people's lives and that has completely undermined their usefulness for learning.

Have we? Or have we allowed grades to get diluted due to perverse incentives? Is the appropriate response to surrender on that or to attempt to realign via better tests? It's pretty important to figure out what the foundational demand is or else we'll just end up reinventing the wheel with new labels. Like with grades, yes the true goal is a complex combination of "how can we motivate students to learn, and measure how well they're learning in a dynamically granular manner, and also measure how well they handle meta tasks related to learning that matter for real world unsupervised activities including continued learning and learning application, and then also convey all that in a high density form that decision makers in different domains can make use of before the student in question has more track record, and also we need to feed that back into how we ourselves are doing with our teaching/methods". That's not an easy problem space, but it's key to recognize that grades became and have continued to be a thing because they've been useful.

As far as I can see, getting back on track is going to require a cultural shift toward grades being a within-class metric of learning only, giving teachers substantially more power in their classrooms to make decisions without fear of challenges and consequences from students, parents, and administrators whose motives have little to no connection to learning.

But students very much actively want to be able to convey their achievements outside of the class. The whole point for most students is linked to the real world, to getting good jobs or other opportunities they desire. "Consequences" are therefore fundamental, touching on various zero-sum games. Of course people are going to care, and some people are going to do better relative to others, and the ones who do worse are going to have incentives to cheat. You can't just side step all that. Like if some colleges stop providing grades as you suggest, game theory out a bit the results of that. Students exit college, and go to apply to some job, and naturally the company wants to know how they did. How do they convey this in a rapid objective fashion? Will companies then preferentially recruit from colleges that do still provide grades? What would the result when it comes to where future students apply to, and the consequences for the colleges without grades they don't? Etc.

I'm not saying alternatives can't be come up with, but at the same time you can't make the basic facts vanish. If one person learns the material really well and is extreme self-motivated and organized and all the various attributes that would result in a legitimate 4.0 GPA, and another one is not, and they both apply to the same job or for the same research grant or the same loan for a startup or whatever, and recruiters place economic value on being able to make special offers to the top 1%/5%/10% and to determine that quickly, how does your approach handle that?

 

[scrimbul]

You have my sympathies.

The embrace of LLMs is just so insanely stupid. I long for the day when the current bubble business model collapses. I’m sure there will be LLMs after that, but they’ll need to be priced beyond casual users.

Fingers crossed. I may be Pollyanna.

Oil prices are going up and never coming back down, I suspect we will all get our wish sooner rather than later when paying for the electricity to waste on token generation forces all the dozen or so big gen-AI players to paywall everything and model evolutions behind pricey tiered subscriptions.

Remedial continuing education courses were always important in community colleges a couple decades ago but now they will be absolutely essential, and offered or placed on top of people who have even less time than before to learn because they fucked it up as a child or were failed between the system and LLMs.

 

[Demosthenes642]

I taught comp sci courses for a few years during my graduate program a few decades ago. This was before smartphones and laptops in the classroom were a thing but the internet was starting to become very very real. Cheating was rampant, even with the higher barrier to entry of having to obtain someone else's work the old fashioned way. Either by request or the ol' over the shoulder borrow method. I was lucky to be teaching at a relatively small and not particularly elite institution where I could give a cheating student a zero without administration coming down on me for "ruining the young adult's future."

Outsourcing one's critical thinking skills is anathema to education and it has downstream impact. Sure the student might get the degree but when they go out into the world, at some point, that critical thinking debt comes due. Innovation and productivity will suffer as will the ability to drive good strategy, it's just going to be about 10 years before we really see the impact.

 

[TMilligan]

This all eerily reminds me of Warhammer 40k. All knowledge of how machines/physics works was lost to man. Since no one knew how they worked anymore then they created a tech religion. All hail the machine gods.

 

[mmiller7]

We are sleepwalking into a world where people are so reliant on LLMs that society cannot function without them. The CEOs of these company know this - in fact, they are counting on it to allow them to raise the rates of LLM product subscriptions for professionals exponentially. Access for students will continue to be low-cost or subsidized by the schools, of course.

Disagree with that - the CEOs want us to think we can depend on them but really its a world where people don't realize they're getting smothered in wrong (and often dangerous) answers.

 

[NedKrist]

Education has value, but apparently students aren't seeing it.

And why is that? Because universities have made it clear that they are profit seeking entities designed to squeeze every last cent out of students and their families while education is treated as an accidental side-effect of the process. The widespread belief in grade inflation means that anything less than an 'A' is a failure. And when they finally do end up with a job (probably not in their field), they'll be told by a cynical middle-manager that almost nothing they learned in university has any relevance to the work they're supposed to do.

There's a perception out there that university is largely a scam*. If that's really what you believe, why wouldn't you cheat?

*I got my degree sometime during the late Iron Age and students (and hiring managers) were already pretty cynical. Still, to me, cheating seemed an obviously self-defeating short-cut. But then I wasn't looking at a permanent lifelong 6-figure debt either.

I appreciate this take on perceptions, but want to add that there are many schools outside the big ones that prioritize student affordability and quality teaching. I've seen the budgets, and there is no profit. Money goes to student support, facilities and paying faculty [edit: and staff] less than they could make in industry. It truly makes me sad that this message doesn't get out more.

 

[Tam-Lin]

A thing I've been thinking about, as someone who has a 9 and a 7 year old daughter: what does this look like to them? Kids already in college now are in an odd place, sort of the opposite of where college kids were 25 years ago, when suddenly we had access to massive amounts of knowledge that was true: they've been using a computer as a source of truth (wikipedia) for most of their lives, and suddenly they have a computer telling them things that are true ~75% of the time, but they still trust it, because computers have been a source of truth their whole life.

For my daughters, AI is anathema. Granted, their teachers are trying to expose them to it, a little, but anything that seems weird, or janky, or false: "That's AI." But they're going to be growing up as the first generation where "the computer told me X" isn't going to immediately mean that X is true. I don't know what that leads to. I do know that I've been reminding them as often as possible is that the point of homework is for them to learn, not to get the right answer.

 

[freaq]

I think the core is that we just need to start testing differently.

back in the day you would get questions to do long division, and with the advent of the calculator everyone’s like oh my God, our kids will not be able to do long division anymore.

So we started asking different math questions.

I think in this case what we need to do is we need to push for oral exams, where there’s no phone allowed to look things up.

perhaps you can bring your own notes, But you can generate by any means you want, but the point is you’ll be able to interpret them and use them.

which is where the learning is.
I just think that written exams are no longer valuable

 

[Plexucra]

The real underlying problem is actually a completely different one. Namely, that students focus on obtaining a degree at all, instead of simply wanting to learn something out of genuine self-interest. Degrees are job-prevention systems. They say very little about whether a person would be well suited to perform a specific job — but at the same time they act as a filter that prevents people from even being considered for it.
The education system in many countries is a learning-prevention system, in that it shifts the joy of learning toward the obligation to obtain a degree. I don’t understand why people don’t finally start addressing these root problems, especially now that large language models are making it so obvious that the previous cat-and-mouse game benefits no one.

 

[Nalyd]

Keep on mind that the US University system makes student take classes and get good grades in subjects that are entirely unrelated to their Major.

While it is good for University to get 4 years of tuition for what should be a 3 year degree, I don't blame students for finding short cuts to get good grades in irrelevant courses.

1) Students are really bad at determining what is irrelevant for them in the future. Something that seems unrelated now becomes the edge that lands you the better job later

2) an education is different than training (though the distinction has been blurred by politics and school administrations). Training prepares you to do one particular kind of job; an education broadens your flexibility and adaptability because we all know the world and job markets are not nearly as certain as they were in the 50s and 60s where a high school degree and some job training set you up for life.

 

[Tactical Finesse]

The real underlying problem is actually a completely different one. Namely, that students focus on obtaining a degree at all, instead of simply wanting to learn something out of genuine self-interest. Degrees are job-prevention systems. They say very little about whether a person would be well suited to perform a specific job — but at the same time they act as a filter that prevents people from even being considered for it.
The education system in many countries is a learning-prevention system, in that it shifts the joy of learning toward the obligation to obtain a degree. I don’t understand why people don’t finally start addressing these root problems, especially now that large language models are making it so obvious that the previous cat-and-mouse game benefits no one.

Well...it is even deeper than that.

Employers use degree-holding as an easy sifter for candidates. Why? Because it is extremely easy for employers to audit if you have one. When you have 400 applicants for a job--sort out everyone without a topical-related degree, even when it doesn't need it. Why? Because less sifting work. Which is why employers over-rely on college diplomas, even when they are completely unnecessary.

For employers--how else do you low-friction easily audit hundreds of candidates for potential competence or work ethic? No one has come up with a viable alternative that works at scale.

So of course kids get degrees. You want to get passed the employer ATS--you "need" one. Of course, with so much degree inflation--the wage-value is diluted and wages are forced down for anyone not on the front-end of that degree's popularity. See early MBA people, or early coder people, or early data scientist majors....as opposed to the tail end, where those credentials became expensive unemployable jokes.

 

[Hoptimist]

I write as if they consider that to be a valuable use of their time. Sure, I think they're wrong, and I don't think it's a judgement most of them are making consciously, but I do think they're making it.

Humans are a social species. Most folks come with a set of basic instincts that tells them to value social connection, and a basic upbringing that tells them to value at least their own group. "Goofing off with friends" is building social connections in the community. Scrolling Tiktok isn't, but it feels like it is when you're doing it.

So, respectfully to your point, yes, I do think they're choosing between a busy schedule of activities that, to them personally, do seem fundamentally worthy.

An alternate view - LLMs isolate students when working outside of class, as they obviate the need to study/collaborate with other students. Students learning from each other is important and instills the value of collaboration - which they will use hopefully both professionally and civically.