Rendered at 13:44:34 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
a57721 1 days ago [-]
I am teaching some math courses, and I see how LLMs disrupted all standard approaches, and I don't know what to do.
I rely on written exams that are open book, but forbid any use of computers and smartphones in class.
The university insists that homework can't be optional, but it lost its meaning. I've tried to explain to my students that it's in their interest to think about home assignments on their own, but the majority will obviously use gen AI, and it's such a waste of time to check and grade LLM output.
I just had an experience where all students were given "sample problems" to try at home and prepare for the written test. Many of them just dumped the document into an LLM, asked it to produce some "exam guide", and showed up with that thing printed out, asking me during the test to explain what LLM output meant.
It seems like the university also has many people pushing for AI use for everything, but I teach basic stuff where the goal is to make students think on their own and digest some fundamental ideas, LLMs can produce perfect solutions, but relying on them is pointless.
professorthread 21 hours ago [-]
I teach math-heavy economics courses. I found that grading homeworks for completeness and pairing the homework due date with an in-class quiz on the problem set material (taken in class the day the problem set is due, after the set is turned in) has reduced the incentive for students to use LLMs to complete the homework. Exam questions closely mirror what's on the homework, and exam grades correlate with students who make legitimate attempts on their homework. I can sus out legit attempts typically by handwriting. No erasing and fast writing means it was copied.
For grade weights, I weight the homework the lowest of any grading category, with in-class quizzes next, and exams scores the highest.
rao-v 6 hours ago [-]
Humm doesn’t this perhaps suggest you should do away with the homework, offer optional practice exercises and double down on in-class testing?
driverdan 19 minutes ago [-]
It sounds like doing the homework is a form of practice exercises for the quizzes.
lazyasciiart 5 hours ago [-]
Perhaps they are also dealing with administration that says things like “The university insists that homework can't be optional”, as in the grandparent comment.
WillAdams 2 hours ago [-]
Seems to me that requiring the homework and having it paired with the quiz works well in terms of cause-effect for the students.
chii 4 hours ago [-]
The university administration needs flexibility rather than draconian rules that dont work.
In any case, even if the uni insist on homework, the teacher can simply mark everyone who hands in their homework 100%. It is impossible for the administration to police it, or if they do, they'd need to hire someone to mark the homework (which conveniently solves the teacher's issue).
grey-area 30 minutes ago [-]
Homework is actually useful practice if you’re not stupid enough to use an LLM to complete it.
i80and 24 minutes ago [-]
I might argue that homework is, for mathematics, where the useful learning happens.
cocoflunchy 5 minutes ago [-]
I wonder if you could see this the other way round? Prior to AI you were limited in your ability to grade handwritten tests.
But now you could have the students take a written exam every other week and use a LLM to grade it?
I'm pretty sure Opus 5.5 has good enough vision capabilities to auto grade with the right prompt.
BeetleB 13 hours ago [-]
> The university insists that homework can't be optional, but it lost its meaning.
The problem is not LLMs, but your university. In my undergrad, almost all the math courses had no required HW. They'd assign it, but you wouldn't turn it in. You'd come to office hours for help on the HW, or ask in class (they'd often dedicate the first 10-15 minutes of each lecture to Q&A).
It was awesome for people like me. No time wasted on neatly presented HWs. I'd do it quickly, verify the answers, and study for the exams.
The flip side was there were many exams (you don't want your grade catered if you do poorly in one exam). A course like calculus could have 4-5 "midterms", and then the final. Often, they'd drop your lowest midterm score so you're allowed one bad day.
trostaft 11 hours ago [-]
Indeed, I've switched over to model similar to this in the course I teach. Weekly optional problem set + in-class quiz every week with one homework problem verbatim + an extension. The homeworks are completely open resource, LLM, friend, sorcery, w/e. I figure that, even if they're given an oracle, they have to internalize the ideas to the point where they can generalize on the follow-up. I'll only really know when the course finishes, but I think this will work out. I want to pair this with an oral midterm and final, but that'll have to wait a semester.
submain 12 hours ago [-]
In my country of origin (not US) that's how it worked starting in high school. Homework was called "suggested problems". The teacher reviewed them in class. No grading. We ended up doing it because that was almost the only way to pass the exams (which, btw, could be up to 50 questions long...)
tha_hnrain 21 hours ago [-]
I think a possible solution is to make homework carry no grade weight, and simply mark it as complete or incomplete, like a presence check.
Homework still provide valuable practice for students who are truly committed to learning. An automated system could provide an initial assessment and feedback. However, students can request human feedback paired with an in-person meeting. Such requests would require a mandatory explanation of what feedback they want or what the issue is. Professors and TAs will therefore only read and answer such requests themselves and will not waste time on AI output. At the same time, the system will still support students who are genuinely interested in learning.
Students who use AI to do their homework are not interested in such feedback anyway, and they will also avoid in-person meetings because it will naturally reveal that they are unprepared and did not do the homework themselves, which is humiliating.
To make this system work, we should present the in-person meetings as collaborative 'working sessions' or group office hours. Encouraging students to sign up in pairs or small groups also reduces individual anxiety.
The main channels for grading remain exams and oral presentations.
gosub100 5 hours ago [-]
In most colleges, if too many students fail the course, the instructor is blamed. So you cannot optimize solely on catching AI cheaters.
lostlogin 13 hours ago [-]
> oral presentations.
These can be written by AI. Or do you mean something more like the defending of a thesis?
lazyasciiart 5 hours ago [-]
Explaining something to an audience, even with notes, is hard to do without actually understanding what you are saying. It’s kind of like a rubber duck.
paul7986 12 hours ago [-]
Committed to learning and going to jump ahead here to say if smart glasses (without cameras or in time with) become the norm and Ai is always right how many will not take the time to learn the information is always right in front of them in their view.
Myself as a student i take my class notes and the subject matter the teacher provides and have Ai create multiple choice quizzes. It's a much quicker way to learn tho not as quickly as AI glasses I mentioned.
procaryote 2 hours ago [-]
Perhaps let them take a practice exam early on, so they can realise they'll fail the real one unless they actually study?
drakonka 1 days ago [-]
I'm taking a year-long sabbatical to study a language at university right now and they're changing their assignment processes to curtail AI use. We have two versions of written assignments - the first to be written by hand in class with just a Swedish-Swedish word book available, and the second version to be typed at home implementing the professor's corrections from v1.
At least in this course students seem to be taking their strict 'no AI' policy quite seriously. Then again this is also the kind of course you do when you want to actually learn the material. It provides a qualification for further higher education in Swedish, so if we don't legitimately learn the stuff there's just no point as we won't manage in future courses.
TFNA 1 days ago [-]
Recently it has been reported that many Chinese students in Moscow are no longer learning Russian during their years at the university (to the dismay of Russian officials who hoped this would be a form of soft power): even for in-person lectures their Bluetooth earbuds will just translate the lecturer's Russian into Chinese in real time. So, perhaps people around the world will start feeling that, thanks to technology, exams testing language skills are a mere formality and can be safely cheated on.
officehero 3 hours ago [-]
Russia is a bad example due to its current international reputation. On your second point, would an example of this be an oral exam where the student has tiny hidden earbuds? Outrageous but you're probably right it's going to get worse.
bananaflag 4 hours ago [-]
> The university insists that homework can't be optional, but it lost its meaning. I've tried to explain to my students that it's in their interest to think about home assignments on their own, but the majority will obviously use gen AI, and it's such a waste of time to check and grade LLM output.
Can't you just not grade it at all? If the student has done their homework, just consider it passed. And then don't count it at all for the final grade.
bambax 1 days ago [-]
There is indeed no point in grading home assignments anymore, but why not randomly ask a student at the start of each class to present their solution on the board?
And do a final exam on oen and paper and nothing else, and an exam that relies more on thinking than rote memorization of formulas (or print the relevant parts of the course on the exam document itself).
hombre_fatal 55 minutes ago [-]
> why not randomly ask a student at the start of each class to present their solution on the board
This makes the class really awful. I had professors with schemes like this, maybe calling on someone's name randomly 1-2 times in class to answer a question they'd just asked, just to ensure everyone is scared into paying attention.
someothherguyy 1 days ago [-]
there probably is no time for 100 students to each do that in a university mathematics lecture, but there are probably ways to make sure students are still learning by restructuring the university into mandatory recitation and labs or something
Aeolun 1 days ago [-]
Having two random people picked each time might cause all of them to at least pay attention.
lazyasciiart 5 hours ago [-]
They said “one person at the start of class”?
someothherguyy 5 hours ago [-]
there aren't 100 university lectures a semester, so there is a small chance you are picked. meaning, good odds to not give a shit or good incentive to not go to lecture if you are behind. also, wasting 100 students' time to watch someone struggle through one gotcha, is more of a punishment for everyone, not something to make students learn.
lostlogin 13 hours ago [-]
I’m wondering if this is the way, but more faculty time is needed for assessing. So education costs go up.
15 hours ago [-]
mppm 4 hours ago [-]
I've had the same experience, and likewise I have no idea what to do.
Or rather, I do have ideas, but they do require substantial restructuring. E.g.:
1) Strict, in-person (no exceptions), electronics-free exams with written and oral portion.
2) Transition to guided independent learning: students are given written materials and recommended exercises to study. They are shown how to use AI assistants productively. Instead of lectures, there are scheduled discussions, where students can ask questions and listen to additional explanations from the lecturer. They can also seek feedback for completed exercises. However, none of this contributes to grades in any way.
With such an incentive realignment, I think it's possible that higher education could be salvaged. But... apart from general institutional inertia, there is also the political angle standing in the way. Everybody needs a degree in the "developed economy", and the above changes are basically the opposite of the nobody left behind policy that enables half the population to attain one...
fhd2 1 days ago [-]
That sounds pretty frustrating. I wonder if it wouldn't be a good approach to, say:
1. Make the exam the only thing that determines the grade (or technically, proves that the student learned enough);
2. Give out test assignments students can do, but don't have to;
3. Offer students for the professor to grade their work if they want to get some feedback on how they're doing.
Would that lead to less time wasted by tutors grading AI generated answers? From where I stand, students should be allowed to prepare for exams any way they like, with or without their professor's help. If they managed to learn, they pass.
psyklic 1 days ago [-]
According to the article, #1 leads to "failure and drop-out rates of 50 to 80%". #2 and #3 might work for organized and motivated students, but graded assignments from other classes will be prioritized by most.
procaryote 2 hours ago [-]
Failing people who didn't learn anything during the course sounds like a desirable outcome
hahn-kev 11 hours ago [-]
Maybe more drop outs is a good thing, we need more skilled labor anyway
isityettime 7 hours ago [-]
In many cases it probably is. But lower graduation rates cuts against the incentives structures in place at many universities and colleges.
gosub100 5 hours ago [-]
Then their parents come in and complain about where the $30k that semester went. This is a consequence of the exponential rise in tuition. You're now in lawsuit territory if you don't deliver the goods.
michaelt 1 days ago [-]
When I was in engineering school ~20 years ago, some lecturers openly said that they gave the assignments just enough credit to be worth doing. Something like 10-15% of your final grade would be from assignments, the rest from the exam. So we didn't blow off the assignments to go drinking.
Of course, this a lot easier for some subjects than others. Subjects like Film Studies relied on exams much less, and assignments much more.
isityettime 7 hours ago [-]
I took some classes like this and they were great. I felt they were much fairer than classes where mere diligence/compliance could masquerade as competency by way of final grades padded by homework scores.
They were disliked by many other students, though, and some people really think university grades should be about things like diligence, compliance, time management, etc., rather than pure subject mastery.
dghlsakjg 15 hours ago [-]
Give a short quiz at the start of the class with one or two problems directly from the homework.
professorthread 21 hours ago [-]
1. is not a terrible idea, if structured correctly, but it doesn't adequately incentive the learning process, which requires continuous practice. For 1, I'd say put 80% of the grade in three to four exams taken over the semester, and allow one midterm exam grade to be dropped or replaced by the final exam grade. The remaining 20% of the grade is to incentivize practicing problems: homework graded based on completeness not correctness, short in-class quizzes randomly sampling a homework problem to incentivize students to legitimately complete the homework that are graded for correctness, and student participation in classtime activities.
Alacart 9 hours ago [-]
> asking me during the test to explain what LLM output meant.
Perhaps (sarcastically) this is the new approach, because this is now the point where they’re forced to struggle and therefore actually learn. An exam every other day, let them ask the questions.
lazyasciiart 5 hours ago [-]
Not sarcastically, this is the flipped classroom approach. Read the examples and watch the lecture on your own time, then come and do the exercises with a teacher available to explain when you realize you didn’t understand anything.
barapa 4 hours ago [-]
If grades matter, the students will cheat.
darreninthenet 1 days ago [-]
Put a mixture of questions in that are impossible to answer but don't tell them, and as usual give them marks for explaining their working. Tell them you've devised the questions so some are "AI resistant" but don't tell them how. They'll never know if the AI has tripped up and hallucinated or given them the right answer. Only those that really understand what's going on will set out some working that shows they were on the right route and then got stuck where they were supposed to.
adrianN 1 days ago [-]
That will either lead to students becoming frustrated after spending inordinate amounts of time on the unsolvable problem or students giving up too early on the solvable problems.
pks016 21 hours ago [-]
> Put a mixture of questions in that are impossible to answer
Students will complain to the admins and waste more of your time.
SamInTheShell 8 hours ago [-]
Why not just make the assignments "reading" based and set an expectation of randomized quizzes that take 5 minutes at the start of class to fill out to get the homework credit? When I say randomized, like multiple versions. You can use LLMs to churn out PDFs to a tailored format and utilize a pool of questions.
xtiansimon 23 hours ago [-]
> “…students were given "sample problems" to try at home….Many…just dumped the document into an LLM…”
I think this development should make all educational programs reconsider the flip-model: homework at school and video lessons at home.
Homework sucks. Learning should be at your own pace (play/pause/rewind).
professorthread 20 hours ago [-]
This is a nice idea, but if you follow the Carnegie model of the number of hours per week it takes to learn a subject per credit hour, which is estimated at something between 6-9 hours total including classtime and out-of-class study per week for a normal-semester 3 credit course, then you can't meet the amount of time students need to be exposed to lecture (explanations, demonstrations, derivations), to do the background readings which provide necessary context that can't always be adequately covered in lecture, and to practice the concepts and tools on their own.
dprkh 15 hours ago [-]
This. I am currently a university student and I wish I could practice in class and not at home.
dghlsakjg 15 hours ago [-]
Office hours, tutoring centres, and self formed study groups are all great options.
dprkh 14 hours ago [-]
For me it's about time management, not about getting math help. I am consistent at attending classes because it's mandatory, but not consistent at finding time to practice math every day. If I could practice during classes, I think that would help me.
dghlsakjg 13 hours ago [-]
That’s an entirely different problem, and a hard one to solve. I would still suggest looking at the tutoring center or groups since there is time accountability there. If you don’t show up it is more consequential than skipping class from a personal point of view.
Footprint0521 13 hours ago [-]
Dude, hop on the Gemini student plan and try Gemini Notebook and Gemini Live. Live is a personal TA, Notebook literally podcasts the readings and what you need to learn
dprkh 13 hours ago [-]
I am not buying this. I use AI to check my work, but it is absolutely terrible at teaching or explaining.
eviks 4 hours ago [-]
Homework is learning at your own pace! You pick the time and place and pace
singpolyma3 15 hours ago [-]
That's a terrible model though. If someone is paying to be taught and you just hand them video lectures... We already have video lectures at home. Why would they pay you? What product are you meaningfully offering in that world?
Wool2662 3 hours ago [-]
You are selling a diploma. Some universities also sell some degree of attestation to the high resilience of their graduates. And of course the space to network. Good teaching is not on the list in my experience.
In general university lectures are horrible (of course there are exceptions). The lecturures oftentimes are required to teach as a condition of research funds, so for many it's not a priority anyways.
On top of that most couldn't care less if their teaching style is didactically sound at all.
wffurr 13 hours ago [-]
Tutoring, office hours, teaching assistants. One of the best classes I took was a "studio style" class where we would work on a project and occasionally the teacher or a TA would interrupt and give a brief lecture or demonstration based on a "teachable moment" that had occurred.
satvikpendem 1 days ago [-]
Just don't make it open note then? Or you provide the only notes or cheat sheet available to them at the start of each test. That's what some math tests I've had had done before and I believe the SAT also does the same thing with formulas in the back of the test to reference.
FuckButtons 11 hours ago [-]
It seems to me that the only realistic option is to have more exams, closed book, and then use gen ai yourself to help reduce the workload of grading. Possibly with some kind of curve assigning more weight later in the semester.
seer 9 hours ago [-]
Hmm people can now do a lot more than they were supposed to be able to without LLMs - shouldn’t we lean into that?
Force people to go all in with “impossible” tasks and see how they do with the power of LLMs. Then spend some significant time teaching people _with_ that solution in place what was actually built and what would have been a better architecture? A … software archeology session if you will.
That’s what most of what people do would be anyway - building PoC fast and then figuring out how to scale them?
viccis 7 hours ago [-]
It's crazy that we went from "we found a significant number of students used AI in our course and gave them a 0" to "yeah they all use it we can't stop it". Basically just accepting widespread cheating, normalizing and accepting the complete abandonment of the first lesson in ethics we are taught and held to in academia.
I know this is absolutely impossible because colleges and universities have headed in completely the opposite direction, but the solution: Literally just expel them. Just like they did when people cheated when I was in college. Don't make it a small punishment so that they turn it into a risk/reward assessment. Colleges are overpopulated anyway; just trim the fat of students who shouldn't be there. Do a follow up oral exam interview about their project and if they can't explain what someone like me spent 30+ hours building and can speak about for 4 hours if asked, then expel them. They can make your coffee at the nearby Starbucks, who cares.
If you don't do this then understand that you are telling your students not using AI that they will need to begin cheating like the others to be able to keep up with the changes. OP mentioned they massively increased the scope of their assignments, which means OP is forcing their students to use AI now.
chemodax 1 hours ago [-]
Agreed. It’s very sad modern society is stopping using ethics and at the same time increase using different firms of policing.
lostlogin 13 hours ago [-]
> I am teaching some math courses, and I see how LLMs disrupted all standard approaches, and I don't know what to do.
I wonder what the financial cost of AI will be to teaching? Does teaching get more expensive? More things need to be supervised, more things need to be presentations or in person?
MaxBarraclough 1 days ago [-]
> Many of them just dumped the document into an LLM, asked it to produce some "exam guide", and showed up with that thing printed out, asking me during the test to explain what LLM output meant.
That student presumably performed poorly on the test. Your approach is working as intended here, isn't it?
Also, is the AI-generated exam guide all that different to a human-written guide they could have downloaded a decade ago?
buckle8017 2 hours ago [-]
Short in person quizzes frequently.
They don't even have to be a significant part of the rubric.
Just ensure the students know what they don't know.
StanislavPetrov 10 hours ago [-]
>I rely on written exams that are open book
How about cutting out the open book so that the students who don't study and to their homework actually fail? When I was in school many years ago "open book" was only used for special ed students and others who couldn't carry a regular academic workload. Homework was assigned, but our grades were determined solely by testing our mastery of the material. If you could master the material without doing the homework or showing up for class on non-testing days, more power to you. College has become more like kindergarten with all of the hand-holding.
isityettime 7 hours ago [-]
When I was in college, whether or not an exam was open book seemed orthogonal to the difficulty of the exam as well as the class's median uncurved score. Allowing the textbook or notes lets you write an exam that might otherwise be unreasonable.
In the classes I took with open note exams where I did well, I barely ended up consulting my notes anyway. The students who didn't study could flip through the textbooks all they wanted, but they lost so much time doing it that they couldn't complete all of the problems on the exam. For the students who studied well, the textbook was way easier to jump around in because they knew it well and knew what they were looking for.
Am exam that is designed to be closed book/closed note where a specific student is allowed some amount of notes as a disability accommodation is just different from one that is designed to be open note/open book from the start. The latter, testing the same material, is generally longer or harder to make better use of the format.
MITSardine 6 hours ago [-]
Open book need not be easier, it entirely depends on the exam.
This was too long ago for me to remember the details of but, just making an example here, we might have had a course on homogeneization of elliptic problems using a model PDE with a div(A grad u) term throughout, and then the exam was about a PDE with a curl(A curl u) and some other added terms.
Notes carry the method and theoretical tools but you then still need to apply them to a new problem.
So, if anything, our open book exams were harder and a better test of understanding and mastery. You couldn’t rote your way into a good grade.
-1 1 days ago [-]
It must be tiresome to be a student these days… with each teacher having different pet theories about AI use and sometimes trying to “trick” students who use AI, or implement some draconian rules about the way homework or exams must be done. It’s tiresome for you, but just imagine getting varying lectures on AI use as you go from class to class as a student. Must be hell
analog31 21 minutes ago [-]
Learning and understanding were always and option, and still are.
fartfeatures 1 days ago [-]
I've always wanted to find out if this would work:
If I knew most of my students were using LLMs I would encourage the rest of them to use them too. I would then change the marking criteria such that if they get a single point wrong that's 20% off the entire grade. 5 mistakes gets you a big fat zero.
LLMs are great but they make mistakes. At this point your students would have had to spend so long checking, re-checking and triple checking the LLM's output that they will have accidentally learn what they need to. They will likely need to cross-reference multiple LLMs and at least have read their output which is likely an upgrade on today.
Or some variation of the above, I'd be interested to hear your thoughts.
foresterre 1 days ago [-]
This sounds like it would come at the cost of the honest students who don't want to use LLM's and train their own understanding.
Reviewing is just not the same as working through a problem yourself. You often can take shortcuts when checking answers for correctness, which generation (with mind or LLM) of the solution can't take. This isn't limited to Math but equally true for programming and many other skills.
fartfeatures 1 days ago [-]
I'd push back slightly on this: I spend more of my time reading / reviewing code than writing it yet I still find reading / reviewing code much more mentally demanding. I don't think I'm alone in that.
The people that do their own work would be at a huge advantage as the LLM(s) can work as reviewers instead of doing the work so there is more chance of catching a mistake before you lose 20%.
On top of that there will be some issues where the LLM is just plain wrong. Being able to work out when that is the case is a hugely useful skill in 2026. It is also something the people who blindly feed their work into an LLM and print its answer without reading it will be unable to discover. Much to their own detriment.
foresterre 1 days ago [-]
I'm not trying to suggest that reviewing (code) can't feel (more) mentally demanding. The thing with at least reviews is that if you have worked through the problem yourself, you will also hit certain walls where stuff didn't work or wasn't as beautiful as you wanted (ie trial and error), and you have to activate your brain to think of other solutions, ie really think a problem through multiple times with active feedback. As a reviewer you don't get the active feedback usually.
Now this i think applies in lesser degree to the comment I was replying to. With maths, and some comp sci problems, you often can check an answer much faster than actually solving it, eg by using simple substitution for simple math problems. That is a useful skill, but a different one from solving the problem yourself. It employs different types of skills.
fartfeatures 1 days ago [-]
I think you might be letting perfect be the enemy of good here. Right now some students are using an LLM to do all of their work. I don't suggest for a moment my method is perfect but I do claim it could be better than the status quo (or at least warrants further investigation in my opinion).
johnnyanmac 6 hours ago [-]
> I spend more of my time reading / reviewing code than writing it yet I still find reading / reviewing code much more mentally demanding. I don't think I'm alone in that.
That's because we write much smaller quantities of code than we read. Not that LoC is a useful metric, but I'd be surprised if I write more than a few hundred lines a code a day in a legacy codebase. Most of the time is spent planning and deliberating over what to write, and perhaps revising code a few times. Meanwhile, understanding what a snippet of code does can easily spiral into having you read thousands of lines of code across several modules.
But that's for work. writing is a stronger learning tool than reading, and we needed to write tens of thousands of lines of code before we got into the door and started reading more than we wrote. It's still preferable for students to do the same and write as much for their homework as possible.
Explaining that paradigm shift from learning to working isn't trivial, and is often poorly done. But it's needed if we ever want people to frame these tools as productivity boosters, and not shortcuts through tedium. Learning by its nature involves some tedium.
thow88737363667 1 days ago [-]
What you would grade with that approach is how well your students can operate a LLM as well as the failure rate of whatever LLM infrastructure the individual student is using. This is usually not the goal of your educational format and hence not what you would grade for - apart from some (meta-)skill courses offered by uni libraries and the like. Those are mostly ungraded, though.
fartfeatures 1 days ago [-]
Anyone that knows the subject matter well enough will spot the LLMs mistakes and correct them. If you don't know the subject matter well enough to spot a mistake then you don't know the subject matter well enough.
sakesun 15 hours ago [-]
In the future, homework and officework will only exists to generate AI providers revenue.
armchairhacker 1 days ago [-]
> such a waste of time to check and grade LLM output
Small nitpick: you teach math, why don't you use an automated grader?
I agree homework should be optional and your college is wrong. Even without AI, for some students it's easy to cheat and maybe for some students it's a waste of time. Maybe you can assign it a very low score, like 5% of the final grade, and/or award the grade based on attempting rather than getting the right answer.
singpolyma3 14 hours ago [-]
Good math grading is looking at things you can't trivially automate. Process, proofs, etc
jmalicki 12 hours ago [-]
Things that, in 2026, are trivial to automate grading of.
singpolyma3 12 hours ago [-]
No?
anal_reactor 1 days ago [-]
> why don't you use an automated grader?
LLM verifies output of another LLM. The circle is complete, and nobody has any doubts that this is just a theatre.
With manual grading, that one student who genuinely tries still has a chance to meaningfully engage.
jmalicki 12 hours ago [-]
Translate the proof to lean and verify it.
If the proof isn't trivially translatable to lean, is it actually a good proof?
armchairhacker 23 hours ago [-]
LLMs have become much better at avoiding hallucinations. Moreover, students can submit grade appeals, which should be reviewed by a human.
pks016 21 hours ago [-]
It would be a waste of time. LLMs have become better but not perfect. Personal experience; it's easier/faster to grade for me than have the LLMs grade and cross-check each one.
throw93038383 1 days ago [-]
> I've tried to explain to my students that it's in their interest
Perhaps some people do not belong at university? It is just a waste of everyone time, to insist 50% of population studies until 25 years old.
dahart 12 hours ago [-]
Who’s insisting that? Where does 25 come from? 4 year degree attainment is just under 40% in the U.S., and most of them finish by 22 or 23.
People are choosing to go to university. It might have something to do with having better job options and statistically higher pay…
MattPalmer1086 1 days ago [-]
25? At least in the UK, most degrees are 3 years long and most people start when they are 18.
TFNA 1 days ago [-]
UK (and the wider Anglo-American system) is different from the continent, where in many countries the MA is the basic degree -- a mere BA will get you nowhere -- and this easily takes 5-6 years to achieve.
ant6n 1 days ago [-]
It's like we're putting everybody on crack and then say they're too weak if they become addicted.
Hendrikto 1 days ago [-]
If you do not want to learn, and take every shortcut available to avoid it, university might not be the path for you.
johnnyanmac 6 hours ago [-]
1. We don't exactly do a great job showing upcoming adults their options and paths in life to begin with.
2. We're also actively giving them less paths in life to begin with. How many high school grads @ 18 will find any job whatsoever without additional school/training, and how else will they afford housing with zero income? The military is the only consistent employer at that age, and the only one who'd provide shelter with the deal.
3. On top of all that, it's in university's best interests to keep 1) and 2) the way they are. Their incentives start at recruitment, and end at enrollment. Past keeping them in college for 4-6 years, they do not care about performance, post-grad job placement, nor societal wellbeing. And part of that is due to a lack of pressure in the recruitment stage to make universities focus on such metrics. Turns out a fancy new gym is a cheaper investment.
threatofrain 2 hours ago [-]
By the time you are 18 and going to college it’s time to reap your lifetime of prep, you’re about to spend a ton of money to specialize. What does it mean to show you all the doors and motivations of life when you’re basically going “all in” on a domain?
Hopefully it doesn’t mean extending high school two years into college, still prepping to make a bet not yet made.
alpaca128 1 hours ago [-]
That lifetime of prep is completely disconnected from actual job & life experience. The kids are shown all the doors but have to choose without first opening any. Aside from the occasional internship they might have done by that point or a glimpse into what their parents are doing, which is very hit or miss.
As long as students ask the teacher what they're going to need understanding of maths for they are not prepared adequately.
fr2029 1 days ago [-]
[dead]
dofm 7 hours ago [-]
I have been feeling for ages like I am too old for the tech industry.
Reading this has made me understand that the recruitment industry will soon have to adopt the default position that anyone who graduated after mid 2025 is likely misrepresenting the depth of their understanding; people who graduated before that are the equivalent of low-background steel, until our processes catch up.
Maybe there’s still room for me.
gregwebs 40 minutes ago [-]
AI is effectively a personal tutor. Personalized tutoring is the best possible way to learn!
The problem is that the student is in charge of the tutor and often tells them to just do their homework for them and not check whether they have learned anything.
LLM providers should formalize a tutoring mode. We should have prompts available to create this mode. This should be the default way students use LLMs.
LLMs may not be good enough to do this now, but I think there is already enough raw intelligence and its a matter of training models to be good tutors, to be experts and specific subjects, and developing software systems for tutoring.
In this model there are few lectures. This is more similar to Montessori but with on demand tutoring. The teacher provides objectives and materials for the students and tutors. The tutoring process can be done during class for teachers to ensure it is working well. There can be larger blocks of office hours for students to do their tutoring during.
As AI gets more intelligent, it will also be capable of doing all the grading. This includes oral exams. Teachers need to proctor the exams in person.
Anyone will be able to tutor themselves on any subject. Proctored examinations to certify knowledge do not need to be very expensive. The value of the school environment will not be to transfer knowledge from a teacher but to have peers working on the same thing and teachers and an environment invested in training the next generation.
tarkin2 14 hours ago [-]
The easiest path is to allow students to cheat on homework and watch them fail in the non-coursework-based exam. Your classes, then, abandon to the un-motivated; with AI enlarging the educational gap.
If you’re forced to attain a pass rate, and can’t watch so many fail, then the certificate becomes worthless in the workplace. Or just offer paid retakes and watch AI make you money.
If you insist on persisting with coursework then a short but heavily-weighted oral part seems to be the only answer, something that will diminish teaching time as a downside, as the article and comments about the flipped classroom touched on.
j45 13 hours ago [-]
The non-coursework based exam can also probably be gamed in the end.
Students could just use the AI to teach/tutor the material to them instead, but the institutions are clutching the pearls too tightly.
stonedivot 11 hours ago [-]
Isn't that the point? To learn? If you're setting a goal that the students must pass, who cares if they get the information from the professor or they teach themselves?
I don't think "learning the material by other means" is the exam being "gamed".
I think the point is most students are cheating and not learning. Where they get the information does seem to matter. Adding friction was needed for most.
chii 4 hours ago [-]
if they use ai to "cheat" the exam by having the ai produce material to learn off, then they literally have just replicated a paid tutor (for cheaper).
Why is that not desirable? As long as the exam does not allow ai usage during it, i don't see the problem.
alpaca128 2 hours ago [-]
Long before LLMs the couple of actually good teachers I had gave very little (if any) homework. Instead their lessons were more dense and exhausting, but once it was over it was over and we otherwise mostly had to prepare for exams.
The education systems were already hopeless before LLMs and now they got caught with their pants down. Of course some problems were unavoidable, but this overreliance on homework was always a problem in my eyes.
garg 12 hours ago [-]
Salman Khan of KhanAcademy many years ago suggested that study should be done at home and homework should be done in-person in class with the help of a teacher. That would solve a lot of the homework problems AI access is causing. But it would require extra effort on the part of the teachers.
Seems the core problem is that many/most students aren't there to learn the material, just to get grades. The class is a complex mechanism to force them to learn material against their will in a validation arms race. The teacher's job becomes something like a military commander. Seems depressing.
professorthread 20 hours ago [-]
This is the way it's always been, sadly. Before AI submissions I had to regularly look for where my materials were posted on sites like Course Hero. Students copy-pasted from Wiki, articles, blogs, etc without attribution. Before that, they were buying or finding solutions books and copying out of there, or hiring people to do their homework/write papers/sit exams for them.
The primary difference now is that the both costs of cheating and ability to verify cheating on the instructor side have decreased dramatically, in tandem. There's also been a kind of mass effect, a cultural change in the attitude of students, where the sheer prevalence of cheating among their peers has made them less likely to feel fear, shame, or guilt over cheating. Students in the past who were caught were likely to confess immediately. Students now are increasingly belligerent and more likely to deny cheating, not just because they believe the professor can't 100% prove it, but also because they believe being caught out for cheating when most of their peers are getting away with it is unfair.
CrimsonRain 1 days ago [-]
This is because we shield children too much from reality. If they had a fire under their ass that made them realize if they don't study well, they will forever be locked into flipping burgers or being homeless, most of them would move their assess.
Don't do anything productive for the society but still get to live a decent life on benefits, job security etc combined with addictive social media have ruined most kids will to study.
And the schools are even worse. When I was in school, teachers encouraged learning things; whether for school or outside, with teacher or self help. Everything was fair game and supported. Debate was normal.
Nowadays the teachers in schools are like a cult. Kids are tuned to reject anything from outside; anything that contradicts teacher is a no go zone. Anything that's "too advanced" compared to what teacher is doing is forbidden. Kids are scared of learning by themselves because the teachers punish it. What do you think these kids will do when they grow up?
MaxBarraclough 1 days ago [-]
> If they had a fire under their ass that made them realize if they don't study well, they will forever be locked into flipping burgers or being homeless, most of them would move their assess.
That strikes me as backward. The higher the expected monetary value of good grades, the more students will prioritise grades over deep learning.
eviks 3 hours ago [-]
The quote is not about grades
parineum 12 hours ago [-]
> This is because we shield children too much from reality. If they had a fire under their ass that made them realize if they don't study well, they will forever be locked into flipping burgers or being homeless, most of them would move their assess.
This seems pretty inaccurate to me at just about every turn.
The fire under their asses is for grades and a degree because they've erroneously been told they need one or else they'll be flipping burgers.
It's such a cliche thing to say but cheating at school cheats yourself of the education you could be getting but, since that's not what most people are there for, they cheat. They are there because they've been told that they need a degree to make a decent living and the degree doesn't technically require learning.
anal_reactor 1 days ago [-]
The least paragraph... has always been the case. My parents told me stories about teachers denying facts not to lose authority. I distinctly recall my teacher pretending that peripheral vision doesn't exist not to lose authority.
dpattila 3 hours ago [-]
> I do not believe that suggesting using low-cost open-weights models would be a fair alternative. (...) We now mention the expectation to use a commercial subscription in our syllabus.
I think structuring a course to be taken using a frontier model subscription is outrageous. If you refine the assignments each semester to be only solvable by a frontier model, your students will always be exposed to the actual pricing of said models. It may be 20 USD/month for now (which is already too much to ask in a syllabus in my opinion), but what if prices increase; where do you draw the line?
What's next -- make the students pay for cloud credits in DevOps courses?
I agree with the need to increase the weight of oral examinations / interviews with TAs though.
Propelloni 1 days ago [-]
Maybe the Germans were right? One in-person test, preferably an interview, at the end of the semester should be enough to motivate the smarter part of the bunch to make sure they understand the issues at hand. The rest is chaff anyway. Don't optimize for chaff.
plorkyeran 22 hours ago [-]
As was sort of offhandedly mentioned in the article, all pre-LLM research into what results in students actually learning the material (as a distinct thing from passing the class) has landed pretty solidly on the side of frequent, low-stakes exams having better outcomes than rare high-stakes exams. The best students learn the material regardless of how you evaluate them; some weaker students will benefit from more frequent nudges to keep up rather than trying to cram at the end, some students simply won't realize they're falling behind without regular feedback, and some students successfully cram at the end and pass the class without retaining much of the material. The final group often expresses a preference for the one-exam-at-the-end model despite being disserved by it if the goal is to learn and not merely pass.
Asooka 7 hours ago [-]
Yeah and look at where this soft approach landed us. We have entitled crybabies voting for Hitler 2.0 because they can't stomach the uncomfortable feeling of seeing a diverse person in their pure white neighbourhoods. Children and young adults need to be hit with the big hard fist of reality repeatedly until they become good people. If some should fall through the cracks, they would not have made for good members of society anyway, so best they fall early and not drag the rest of us with them.
djeastm 13 hours ago [-]
They pay as much tuition as the wheat and the schools want them to keep paying said tuition.
YmiYugy 1 days ago [-]
I had that a variation of that experience. This was mostly in the pre-AI days.
Homework assignments were mandatory but only counted towards exam admission not the final grade.
Much better for motivation.
tgv 1 days ago [-]
When I was a student, nobody cared if you did your home work. The teacher would go through the home work in the next class, and if you didn't do it, you would just lose the opportunity.
Then there was the written exam at the end of each course, which constituted 100% of the grade. The only reason that got abolished was because it made too many students fail, and the board hates that. So that became a race to the bottom, and apparently, we can still go lower.
It's reaching levels where most uni diplomas don't matter any longer. The only thing that will then matter is your network. That's the end of social mobility.
armchairhacker 1 days ago [-]
One test is too stressful and may be affected by bad luck. I'd recommend many small quizzes. You can use LLMs to grade.
lostlogin 13 hours ago [-]
> You can use LLMs to grade.
The article gives examples of LLMs being wrong. An AI grading of an AI submission would presumably give a good mark, to a wrong answer.
tefkah 1 days ago [-]
on the off chance you’re being sincere: the article mentions that they allow free retakes of the oral exam
anal_reactor 1 days ago [-]
It's terrifying that some people actually do think this. Not you obviously, but I had enough conversations with other people.
ares623 15 hours ago [-]
"Don't optimize for chaff." is being chaff a permanent state of a child or person?
I think I understand what you mean though, and I kind of agree. I just don't like the use of the word chaff, like something you would discard permanently.
I'm pretty sure I was "chaff" at some point (maybe multiple) in my life growing up.
andrewflnr 14 hours ago [-]
It may not be permanent, but by the time someone goes to university it's not the instructor's job to fix them.
anal_reactor 1 days ago [-]
I agree.
Barrin92 14 hours ago [-]
Funny to see this comment, that's exactly how studying math in Cologne was. On the very first day one of our profs said:
"This is math, you can find all solutions to the course problems in old material since I've been teaching here for 20 years, but know this you're cheating yourself and at the end of the semester you're taking a three hour test on pen and paper so good luck"
I see zero reason why AI should impact teaching. You can grade exercises as information to the students as to how well they're doing, but simply have an exam at the end of the course. Needlessly to say like 60-70% flunked through Analysis I and Linear Algebra. But they're adults and they're there voluntarily, so why is this an issue
stonedivot 11 hours ago [-]
Honest question: doesn't this highlight how mundane and rote teaching has become? It's been distilled to the point that the lowest of the lowest common denominators can pass. The kids going through it are forced to learn things they will never use, unless it's going to be used in the more advanced class of the thing they'll never use until they graduate.
My hope is that this is a wake up call to teachers / professors to find a better way to evaluate students. To determine who has actually come away with a real understanding, vs those who were running around copying the assignments off of their peers anyway. These same people are now just skipping the peers and going straight to the all knowing oracle. It's the same as it ever was, degree mills.
MaxBarraclough 27 minutes ago [-]
> doesn't this highlight how mundane and rote teaching has become
The example questions given in the article do not assess wrote learning.
> It's been distilled to the point that the lowest of the lowest common denominators can pass.
Have pass rates been increasing? Unless I missed it, the article doesn't say so.
> forced to learn things they will never use
A computer science degree is not a coding academy, it's a basic grounding in an area of study. A computer science graduate should have some understanding of theoretical computer science and of computer architecture, say, even if they're unlikely to apply these topics directly in their careers.
> My hope is that this is a wake up call to teachers / professors to find a better way to evaluate students. To determine who has actually come away with a real understanding, vs those who were running around copying the assignments off of their peers anyway.
Academics are already aware of the importance of fair assessment, but it isn't easy, especially with LLMs in the mix.
> These same people are now just skipping the peers and going straight to the all knowing oracle. It's the same as it ever was, degree mills.
Students that cheat, and degree mills, are two different things.
pards 2 hours ago [-]
> Minor note: We observed that some students used AI during live discussions over Zoom (e.g. Cluely)
I frequently see this in remote interviews for senior developers. These are conversational interviews, not coding tests, but some candidates seem incapable of having a conversation without AI.
I can't imagine how these candidates would be able to meaningfully contribute to a whiteboard design discussion with real people.
For a previous client, we resorted to a simple in-person pen-and-paper screening questionnaire with 5 very basic questions. Many candidates declined to attend, and most of those that attended failed dismally.
mtrx 1 days ago [-]
I was at EduLearn (a major edtech conference) this year and talked to many professors exactly about this. A lot of them are pulling work back into the classroom, but keep running into the same problems:
- some skills only develop through the actual writing (similar to how you can't learn to code just by watching YouTube videos)
- in-class assignments are not always scalable (and students end up using AI anyway)
- students aren't motivated enough to participate actively
I think looking at the writing process is a more realistic approach, so I built Turingo (https://www.turingo.net). It replays how a Google Doc was written, highlights unusual edits (large pastes) and tracks how that text changed afterwards. Surprisingly, several professors reported that students often underestimate how much LLM-generated text is left in their final drafts, so just seeing the replay changes the conversation.
It's not meant to be a verdict machine like the existing AI detectors (another big problem in education), more of a starting point for the kind of conversations the author's TAs are having. In fact, we're also beta testing a feature that suggests a few questions for the professor to ask based on the replay.
And yes, it's not perfect and autotypers do exist, but they're dumber than you'd expect and nowhere close to mimicking a real human (I hope to address that soon).
1 days ago [-]
somewhereoutth 1 days ago [-]
It sad that it has come to this, but your approach is correct - it is the act of writing (code or prose) that is important, not the finished output itself.
Interestingly, this applies not just for learning, but also to a great extent in production/work contexts too.
helsinkiandrew 1 days ago [-]
> I now shifted to in-person interactions with a TA. After every assignment, each student needs to schedule a 15-minute meeting with a TA to answer a couple of questions in a live conversation
Isn't the answer the 'flipped classroom' - students are assigned reading/studying to do before a class and the class becomes more interactive, answering questions/discussing topics/solving problems etc depending on subject. Of course, this is more expensive than a hall with 400 people listening to a lecture with a couple of tests/essays.
This only works if you either have very intrinsically motivated students or you verify that really everyone has actually read the assigned materials before class. Otherwise, some people will turn up unprepared, you will have to explain stuff they should have gotten from the reading assignment, other students will be annoyed, students will stop preparing for class, vicious cycle.
chrismorgan 1 days ago [-]
And then you fail them. I suspect an institute genuinely empowered to do so, where the expectations were declared up front and followed through on consistently, would fare well, after the dust settled.
Espressosaurus 1 days ago [-]
As far as I can tell (based on university and high school level instruction forums) institutions increasingly are not backing the teachers that want to let kids that have earned that F fail the students.
I've heard of places where the minimum points for an assignment is 50% even if you never turn it in which is insanity.
raphman 22 hours ago [-]
FWIW, this isn't a problem at my German university.
Espressosaurus 21 hours ago [-]
This is for American schools, unfortunately. And it's not a good development.
chii 3 hours ago [-]
It stems from the idea that "no child left behind" is a goal. It shouldn't have ever been.
raphman 22 hours ago [-]
I don't fail them usually - I expose them; and once I asked a student to leave the room because they were blatantly unprepared despite me warning them before.
My main point is: flipped classroom works great in some
contexts but is no panacea and requires setting clear expectations.
Aside: as English is not my native language: "you fail them" can mean both "you give them a failing grade" and "you fail to provide the support they need", right?
chrismorgan 20 hours ago [-]
> "you fail them" can mean both "you give them a failing grade" and "you fail to provide the support they need", right?
Yes, but as written the latter would not be a reasonable parse. For that sense: “you have failed them”, sure; “you are failing them”, sure; even “in acting so, you fail them”; but “and then you fail them” as an entire sentence would be very weird. The emphasis I placed on “fail” also doesn’t feel particularly compatible with that sense of the word.
officehero 8 hours ago [-]
As a former flipped classroom student I can attest to the effectiveness of "exposure". It requires some skill to pull off: the first time someone is exposed it should be very lenient and then get gradually harsher. The risk of embarrassment in front of the class can be a very strong motivator if utilized right.
Aeolun 1 days ago [-]
[dead]
foresterre 1 days ago [-]
I've been in uni classes which used this concept (about 10 years ago though). It also requires that content is given in the right size and complexity each week, otherwise there will also be people coming to class who did do their homework but didn't understand it properly and just like the people who didn't read it need full explanations.
I also felt having multiple explanations (in a regular class where you could also read the book in advance) from regular lecturers helped me in gaining understanding because having something explained in multiple ways made sure there would be at least one way, or a combination, which made it 'click'.
I always liked the regular lectures where you would be required to do some pre class reading and a smaller set of exercises the most, as you would be engaged in the class itself and would be getting repetition with different explanations.
cheesecakegood 8 hours ago [-]
Flipped classrooms are like, the quintessential siren song for teachers. EVERY teacher dreams of a day when they can focus on what they're best at (teaching deep topics, getting 'aha' moments, making connections, asking questions, and so on) and do less of what they dislike. But sadly this is just not how things work.
Flipped classrooms have incredibly weak causal evidence. However they are popular in spite of this - perhaps because of this if you're cynical - as the educational research field is pretty questionable in its standards and the profit motive is a strong thing lurking in the background. Wikipedia has a big warning on the article for a reason, but even then it has serious issues.
"Active learning" is a thing (insofar as it has meaning rather than be a useless buzzword catch-all) that does work. But the "flip" specifically does not, or rather, sometimes a mix of extra effort on behalf of the teacher and other loosely related effects occasionally can make a modest impact that makes up for its failings.
Most of the people pushing flipped classrooms are the same people who think that students learn best when they "discover" concepts on their own, maybe with a bit of guidance. This is the other, more serious, lie in education right now, akin to the whole idea that you didn't need to teach phonics to kids when they learn to read.
pks016 20 hours ago [-]
One of my colleagues tried it for a few terms. Unfortunately, it didn't work well. Most students don't care about learning, they are here for grades :(
lostlogin 13 hours ago [-]
> Most students don't care about learning, they are here for grades
It’s a virtuous circle though, as many universities are there for the money, so it’s a beautiful exchange.
eviks 4 hours ago [-]
> more importantly I do think that students need to learn responsible use of these technologies anyway.
But you're not doing that, you've just given up and allowed any use!
> How to use AI coding agents effectively is not a learning goal
Exactly
> violate evidence-based best pedagogy practices, and I made them anyway.
Of course you have, when have educators cared much about best practices
> We have since built infrastructure to autograde code and written reports of homework with LLMs (institutionally approved LLMs)
Further reducing the point of your own existence
> Debriefing after such a failure is a good learning opportunity to talk about automation bias and different forms of human oversight (and how this is really hard in practice).
It would be if there were any learning, but since the incentives are for more AI use and less learning, there will be no application of the supposed learning, so you'll continue to have those 80% fails
bartvk 2 hours ago [-]
> Of course you have, when have educators cared much about best practices
Why do you say so?
This is not at all my experience, as a recent teacher it strikes me how much the teaching methods are supported by science.
nfrankel 1 days ago [-]
I have a fairly long experience of both teaching and grading, even though I stopped a couple of years ago. After a couple of years of trial-and-error, I think I found the "right" way:
* Alternate between theory and practice in the course of a single 2-hours slot. Theory is necessary, but the attention span of people is very limited.
* Grade either a homework assignment in the form of a real project with specifications. You give the assignment half-semester, so that students who start early can ask questions. Otherwise, grade a in-session exam, with every resource available, including the course and the whole Internet. However, you don't assess the knowledge, but if the student is able to apply their knowledge to the different tasks at hand.
These approaches are now completely moot in the age of AI.
I was surprised the first time a colleague asked me to do an oral exam. I thought it was a burden on people with poor social interaction skills. It went surprisingly well: you could see in a couple of minutes if the student had integrated the concepts of the course, even if they were shy. Now I wonder if it's the way to go. There's one caveat, though: it doesn't scale.
leoedin 1 days ago [-]
I was thinking that one of the great ironies of artificial intelligence is that it will make good teaching even more labour intensive. Those kids that can afford the one to one tuition required to push past our natural inclination to be lazy will learn things, while everyone else will succumb to outsourcing all their thinking.
There’s no way I’d have learnt everything I know now about software in today’s environment. So much of software development - debugging techniques, architectural decisions, structuring data etc is learned through trial and error. As the models get better is becoming easier and easier to just not look too closely at their output. I suspect it’s human nature.
BeetleB 13 hours ago [-]
> Alternate between theory and practice in the course of a single 2-hours slot.
2 hours is a lot! In my universities, classes were either 50 minutes, or 70 minutes. I could handle 50 minute ones just fine, but would often have trouble with challenging courses that had 70 minutes. Particularly for stuff like math, I hadn't properly digested the material in the first half, and now he's already proving theorems with that material in the 2nd half. I can follow the logic, but the ideal is when you can digest the prior material before being exposed to theorems relying on them.
Even when alternating theory/application like you say, the amount of attention I'd give to the application in the 70 minute class was less than the 50 minute one. I could follow, but I wouldn't think critically.
universa1 1 days ago [-]
The oral exams scale better than one might expect, at least if you're not just doing multiple choice exams and you have to actually grade the exam. The time spent on grading the exam could have also been spent on an oral exam... For bachelor level it's usually 15-20min (usually quicker for the well prepared students), and for master courses it is 30min.
Usually you know after a few minutes if it's going to be a fail, and then otherwise you only need to figure out where on the passing scale the student ends up.
lukehandcool 1 days ago [-]
The country of Argentina, despite its crippling debt, is able to provide free university education where the majority of classes for most majors are graded with an oral exam.
They've shown that it is entirely possible to scale a system like that, as long as the society sincerely values the role of an educator.
Propelloni 1 days ago [-]
Oral exams don't scale at the beginning with hundreds of students per course, agreed. So make them in-person written exams, German universities managed that just fine a few decades ago. Oral exams were reserved to advance from Grundstudium (basic) to Hauptstudium (main).
In my very limited experience (I have German major degrees in Informatik and Philosophy, I was a TA for Logic in Philosophy) the more advanced the topic the thinner the attendance anyway.
naveen99 1 days ago [-]
They do scale if the examiner is an LLM.
lostlogin 13 hours ago [-]
The article mentions study content that AI has got confidently wrong. That’s going to be some crap markings.
CrimsonRain 1 days ago [-]
It scales just fine. Use an AI to do oral exam (recorded). Flag the problematic orals and reevaluate the scripts/recorded session.
Finally, if a student feels their result is unjustified, reevaluate the recorded session.
Aeolun 1 days ago [-]
Use more AI to solve the problem of AI?
smugglerFlynn 1 days ago [-]
I cannot shake the feeling that we are chasing the wrong goose.
The end goal of education is to help you understand how things work, with problem-solving being a means of developing and applying that understanding. LLMs shortcut the former and turn the latter into a pointless arms race that we cannot win.
I loved the example of increasing the complexity of learning projects, but I still feel like one question is missing: are people really trying to learn the same thing we were trying to teach them in the first place?
tra3 8 hours ago [-]
> Similarly, I always provided a safety net where students can make mistakes and resubmit a limited number of assignments to regain lost points (a core recommendation of specifications grading and grading for equity to focus on learning outcomes, not the process), but we felt that this process was abused with AI: first submit a generated assignment solution without thinking and only look at the issues raised in grading for a resubmission (the typical story of externalizing the cost of AI use).
This is how we do PRs now.
charlieyu1 1 days ago [-]
20 years ago my department head was saying homework was pretty pointless because everyone was copying each other. They assigned a token percentage for homework so at least the students were submitting homework.
zkldi 3 hours ago [-]
I find it offensive that your blog is named after thelastpsychiatrist when the writing is not even close in quality. Come on man, I expected you to start swinging. Are you building an agent tunnel? Get outta here.
meken 11 hours ago [-]
> For example, the evidence favors frequent low-stakes assessments with feedback (e.g., homework, quizzes) over few high-stakes ones (e.g., exams) – but AI is undermining practice in low-stakes settings and pushing us more toward exams.
Why would AI undermine quizzes? I took an intro to Japanese class where we would have a quiz at the beginning of each class which I quite enjoyed. I can’t see how AI would change the utility of that.
chanux 7 hours ago [-]
Incredibly insightful and makes me feel bad for those who have to put a lot more work to just maintain basic standards (Teachers, TAs etc.).
It was also interesting to see that everyone is now forced to use LLMs to do the assignments. I wonder if students are using any kind of subsidised access to LLMs here.
thow88737363667 1 days ago [-]
I teach mostly smaller courses but my approach to grading has been the same for years: A combination of a semester-long group project (ending in a paper, a prototype and a group presentation) with individual oral exams about aspects of the project and of the lecture. The lecture is flanked by optional TA sessions as well as access to labs which, for most hours of the day, have some experienced/teaching people being around (this results in a lot of ad-hoc learning situations and a general diffusion of knowledge and practices).
This gives me/us a lot of angles to understand what progress has been made throughout the semester as well as to consider and react to individual differences between students. In my book, it’s also resistant to faking it convincingly throughout all modalities. However, scaling that approach to big entry level courses would require more money for teaching than any research group in Europe I’ve ever seen has available.
Aeolun 1 days ago [-]
I dunno, if I had to teach students now I think I’d just force them to stay in the lecture hall without internet for the hour I need them to work on their assignments that week. I just can’t see any other reasonable way to do this. Everyone leaves their phones at the door in some designated locker or something.
procaryote 2 hours ago [-]
Why enforce effort rather than outcome? Just have exams be without phones.
If someone manages to learn without doing the homework, or finds some way to learn better using an LLM, fine. The exam will show that.
Peacefulz 21 hours ago [-]
> "Everyone leaves their phones at the door"...
— watches, calculators, discrete technological devices. I've seen kids encode text as patterns of dashes on their desks, sharing the cipher solution with their desk mates in other periods.
The problem isn't skirting the effort. The problem is systemic. Why does the system reward taking shortcuts? I think it's because most areas are focused on short horizon tasks, min-maxing effort:output. It's not a problem with those technical fields. It's a problem with the current crop of the C-Suite across the entire occupational plane. They're short sighted.
N_Lens 1 days ago [-]
Thanks for sharing. Interesting to see how teachers are adapting to LLMs. I agree that trying to ban AI use is futile.
saimiam 14 hours ago [-]
Think of education as a wishing well. There is a wishing well for math, law, dentistry, baking, and so on.
Once you have a wishing well - i.e. you own an eduction in baking or dentistry or law or math or whatever - you can ask it anything and it will grant you your wish. If others want something from your well, they have to come and ask you. What you give them has your name attached to it. So, your reputation rides on the quality of what your wishing well produces.
Prior to AI, universities chose which wishing well they would specialize in building and offered to share that knowledge with students.
In the age of AI, the creation of personal wishing wells has become trivial but universities still think they are in the business of teaching people how to create wishing wells.
Ultimately, the only truth left is for owners of the wishing well to start offering the product of their wishing well to the market and attach their name to it. If there is a market for their product, great. If not, go make another wishing well.
Education used to be pets. Now, it is livestock. Universities are not willing to accept this shift in mindset.
sfink 11 hours ago [-]
Heh, dumb idea of the day: do a "flipped classroom" in a different sense -- instead of a student taking a quiz or exam, you provide a clueless AI and the student has to train it on the subject at hand, and then the AI takes the test.
Lest anyone think this is a brilliant idea: an LLM would likely be much better at doing the ~~RLHF~~ RLMF, so you haven't actually gained anything. But I feel like there may still be the kernel of a good idea somewhere in here.
Perhaps your assignment is to iteratively train the AI, which is pre-prompted to seek out every reasonably possible failure mode when it generates solutions. So the learning mechanism is to train an adversarial AI on the subject matter as a way of learning it yourself. This does not solve the problem of cheating with an LLM, it's an alternative pedagogical approach.
pks016 20 hours ago [-]
I don't like it. Now, the TAs have high cognitive load. Earlier, you only need to know the material of the course. But now, you're forced to navigate AI output. And you got to keep with all new model developments.
generalizations 10 hours ago [-]
Of course what's interesting is that when viewed from the perspective of 'I have a kid and I'm willing to orchestrate their education myself', AI means the kid can get an education unheard-of even ten years ago. Personal, genius-level tutor, expert in any subject in which the kid can muster a question.
The only people really having a crisis here are teachers who want to feel relevant.
Honali 1 days ago [-]
The most interesting part here is that AI isn't replacing the learning goals, it's replacing the evidence instructors used to measure them
nurettin 1 days ago [-]
We still have developers because we can't outsource responsibility to AI. If shit breaks, we still need someone to fix it. Maybe you think that can also be automated. I have a counter-example. AI will rarely use the existing state when solving a bug. It will tack on it's own event queue, it will make up it's own messaging system, it is a cancerous growth. And this observation is based on both opus5.5 and gpt5.6-astra. The results are way past magical at this point, but you still have to weigh in pretty heavily for orderly code growth where solutions are succint, focused and preserve the existing structure. They are still noobs at fixing/adding stuff.
jdw64 1 days ago [-]
This makes me think of exams — I always did well on open-book exams, but I didn't do well on exams where I couldn't look at the book. Whether it's because of ADHD or just a lack of memorization ability.
Similarly, I did well in interviews where I could use the internet, but I didn't do well in interviews where I couldn't use the internet.
Everyone has a testing format they're good at. I'm curious about how this kind of thing is evaluated, and whether there are any papers related to this.
sambapa 1 days ago [-]
Why teach someone who doesn't want to learn?
copperx 21 hours ago [-]
Why should we make someone work who doesn't want to work?
derwiki 19 hours ago [-]
We don’t
jan_m_savage 1 days ago [-]
very simple. All exams to be paper and pen.
flimflamm 1 days ago [-]
Thinking too small. The interactive testing needs to be done by AI. Then the transcript is reviewed by the teacher.
parineum 12 hours ago [-]
If we're talking higher ed., expell students that use LLMs to do their assignments for them. They obviously aren't there to learn and don't belong there.
whattheheckheck 15 hours ago [-]
Make all reports and assignments open source and public. Their credibility as students and professionals will be on the line. It will be obvious who puts in the effort and who doesn't
saimiam 15 hours ago [-]
building a community around open source is hard. building one that analyses, comments, and votes on contributions is harder. who is going to read and react to the open sourced content that students submit?
egl2020 13 hours ago [-]
This seems much better than hand-wringing and woe-is-me that we usually see. Good job.
BrenBarn 1 days ago [-]
In my experience a significant proportion of "evidence-based best pedagogy practices" is hogwash. I'm pretty anti-AI, but going against education-school dogma isn't a reason to avoid AI.
bartvk 2 hours ago [-]
Why do you say so? What is your experience?
I recently did a course on teaching bachelor-level education, and had to study a couple of books. They all seemed to quote their sources and I got the impression that they're rooted in science.
Madmallard 5 hours ago [-]
Is society really trending in the direction where learning to do homework assignments is meaningful?
I think society is moving back toward nature at rapid pace. More adversarial in general, and advantage taking behavior is the way to survive if you don't have a solid community. Society will be culled to a much smaller percentage of people than are present today, and it probably won't be pretty. Smaller communities instead of larger communities, with self-sufficient systems in place and a lot of warfare...
Don't see how AI overlords controlling everything will work out. If no one can buy the food from the large producers, then they will dissolve too. Supply chain will fail at some point in the next couple years maybe? Diesel is already outpacing wages for truckers.
The people with all the money are positioning themselves to reap the assets from others to give themselves as much leverage as possible.
At this point I don't really see how it could go any other way.
I don't know. I'm trying to figure out what's going to happen, but it doesn't really seem like anything good is going to happen because the trend is wanton disregard for morality in exchange for stealing wealth.
Makes me really sad. I'm also on fixed income so I worry quite a lot about my future.
bitwize 1 days ago [-]
School homework is one of those things which, like the OnlyFans economy, I hope AI utterly destroys, despite my general detestation for the corrosive effects of AI on society.
varjag 1 days ago [-]
You're not learning anything if you don't put in your hours. Maybe you're fine with that but make sure you realize what you're getting into.
joseluispino 33 minutes ago [-]
[flagged]
NeoByte 5 hours ago [-]
[flagged]
Toprogos 10 hours ago [-]
[flagged]
hieulouisdev 8 hours ago [-]
[dead]
colenikol2 1 days ago [-]
[dead]
otabdeveloper4 1 days ago [-]
Just grade the process, not the result. It's literally that simple.
AI can talk the talk, but it can't walk the walk.
A simple "explain why this is the best solution" will make it shit the bed 10 out 10 times.
DonHopkins 1 days ago [-]
[flagged]
vjk800 1 days ago [-]
It's funny how people talk about "teaching" getting more difficult with LLMs.
Clearly teaching got a lot better, since we now have a new teaching tool at our use, the LLMs. It's evaluating that got harder, because students can use the same tool to cheat in traditional forms of evaluation.
If "teaching the students" means the same as "evaluating the students" to you, maybe you shouldn't teach in the first place.
adrianN 1 days ago [-]
Teaching is imo mostly about techniques that battle laziness. Teaching intrinsically motivated intelligent students has always been trivial. All progress in education is about scaling it to work for people who only reluctantly engage with the material.
cheesecakegood 7 hours ago [-]
Right, and part of that is nurturing student interest and buy-in, because it can increase. Traditionally smart teaching and the right structure could result in some percentage of reluctant learners being drawn into good habits and better retention. I think the subtle part of the AI in education crisis is how it basically has smothered that "nurturing" in the cradle, right from the get-go, so it's a bit of a cold-start situation in many classes.
YmiYugy 1 days ago [-]
Teaching and evaluation are of course not the same thing, but evaluation is an important part of teaching.
There is a mismatch between students’ overall desire to learn things and the moment-to-moment experience that learning is hard.
Even devoid of all the economic and social pressures that grades impose students would compromise their learning experience.
Ever regretted looking up the solution to a puzzle in a video game or sneaked a peak at the crossword puzzle solutions?
One of the hardest part of teaching is keeping students from self-sabotage.
Evaluation helps with that.
Beyond that evaluation is of course also useful just as feedback.
chrismorgan 1 days ago [-]
What you’re calling evaluating is a part of the learning process for the student, not just about confirming that the student knows the thing. LLMs are actively undermining learning in this way. There was useful friction which has been removed.
tgv 1 days ago [-]
Without grading, students don't learn. It's as simple as that.
thow88737363667 1 days ago [-]
That outlook is way too bleak. As a person who has taught at universities for fifteen years and has run lots of open educational formats and infrastructure with other public institutions and citizen groups, my experience is the opposite. If anything, the desire to learn and curiosity are among the strongest common traits humans possess. Grades are just a tool and a proxy that most people working in formalized educational settings have to work with at specific points.
psyklic 1 days ago [-]
Motivation can come from many sources, not only grading, e.g. individual encouragement, exciting applications, day-to-day relevance, interest shown by peers. Also students can just be intrinsically motivated.
mantas 20 hours ago [-]
It may come in different styles. But what do you do with students who don’t have motivation beyond grading/punishment? Some people just do not have motivation unless somebody standing by them with a stick.
defrost 1 days ago [-]
Pretty strong opinion that overlooks all those students that study or attend courses specifically to learn.
mantas 1 days ago [-]
It’s not just evaluating. Students used to learn skills while doing an assignment. LLMing through assignments teaches copy-paste at best.
It’s like copy-paste Wikipedia presentations 2 decades ago. People used to learn things doing research for a project. Then suddenly it became copying off Wikipedia and people not even pre-reading what they copied into PPT.
singpolyma3 15 hours ago [-]
> For example, the evidence favors frequent low-stakes assessments with feedback (e.g., homework, quizzes) over few high-stakes ones (e.g., exams) – but AI is undermining practice in low-stakes settings and pushing us more toward exams.
Only because you're obsessed with being able to assign grades and fail cheaters.
two_handfuls 14 hours ago [-]
Or possibly because quizzes and homework are a good way to learn the material, but if students have the AI do it instead they won't learn.
No obsession required.
OroPla 3 hours ago [-]
What is the point of learning something an AI can already do better than you? There are better methods of brain training.
The situation reminds me of what my teachers used to tell me. "Learning how to look up things in a book will be a skill you'll need all your live and you won't carry a calculator with you at all times when you are grown up." Both assumptions were proven wrong before I was even out of school.
If the only remaining value of education is "to become a well rounded person" then why waste your time on it? It used to be you went to university because that was the only place to get knowledge from. Which is also why it was dominated by the rich and powerful. These days university feels a lot like paper money. Everyone just pretends it has a real value and because everyone agrees, that's where the value comes from.
singpolyma3 14 hours ago [-]
The students who want to learn aren't having the AI do it for them. And they will learn better in the evidence based system.
The ones who don't want to learn will cheat themselves of an education. This is their own problem and we shouldn't punish the actual students to catch these others.
yjftsjthsd-h 14 hours ago [-]
Failing to catch the cheaters does punish the honest students, who in your approach get a worse GPA.
singpolyma3 14 hours ago [-]
This is exactly what I said. The obsession with grades (GPA) over education is the cause here. Intentionally making the education worse in order to protect the grading system.
I rely on written exams that are open book, but forbid any use of computers and smartphones in class.
The university insists that homework can't be optional, but it lost its meaning. I've tried to explain to my students that it's in their interest to think about home assignments on their own, but the majority will obviously use gen AI, and it's such a waste of time to check and grade LLM output.
I just had an experience where all students were given "sample problems" to try at home and prepare for the written test. Many of them just dumped the document into an LLM, asked it to produce some "exam guide", and showed up with that thing printed out, asking me during the test to explain what LLM output meant.
It seems like the university also has many people pushing for AI use for everything, but I teach basic stuff where the goal is to make students think on their own and digest some fundamental ideas, LLMs can produce perfect solutions, but relying on them is pointless.
For grade weights, I weight the homework the lowest of any grading category, with in-class quizzes next, and exams scores the highest.
In any case, even if the uni insist on homework, the teacher can simply mark everyone who hands in their homework 100%. It is impossible for the administration to police it, or if they do, they'd need to hire someone to mark the homework (which conveniently solves the teacher's issue).
But now you could have the students take a written exam every other week and use a LLM to grade it?
I'm pretty sure Opus 5.5 has good enough vision capabilities to auto grade with the right prompt.
The problem is not LLMs, but your university. In my undergrad, almost all the math courses had no required HW. They'd assign it, but you wouldn't turn it in. You'd come to office hours for help on the HW, or ask in class (they'd often dedicate the first 10-15 minutes of each lecture to Q&A).
It was awesome for people like me. No time wasted on neatly presented HWs. I'd do it quickly, verify the answers, and study for the exams.
The flip side was there were many exams (you don't want your grade catered if you do poorly in one exam). A course like calculus could have 4-5 "midterms", and then the final. Often, they'd drop your lowest midterm score so you're allowed one bad day.
Homework still provide valuable practice for students who are truly committed to learning. An automated system could provide an initial assessment and feedback. However, students can request human feedback paired with an in-person meeting. Such requests would require a mandatory explanation of what feedback they want or what the issue is. Professors and TAs will therefore only read and answer such requests themselves and will not waste time on AI output. At the same time, the system will still support students who are genuinely interested in learning.
Students who use AI to do their homework are not interested in such feedback anyway, and they will also avoid in-person meetings because it will naturally reveal that they are unprepared and did not do the homework themselves, which is humiliating.
To make this system work, we should present the in-person meetings as collaborative 'working sessions' or group office hours. Encouraging students to sign up in pairs or small groups also reduces individual anxiety.
The main channels for grading remain exams and oral presentations.
These can be written by AI. Or do you mean something more like the defending of a thesis?
Myself as a student i take my class notes and the subject matter the teacher provides and have Ai create multiple choice quizzes. It's a much quicker way to learn tho not as quickly as AI glasses I mentioned.
At least in this course students seem to be taking their strict 'no AI' policy quite seriously. Then again this is also the kind of course you do when you want to actually learn the material. It provides a qualification for further higher education in Swedish, so if we don't legitimately learn the stuff there's just no point as we won't manage in future courses.
Can't you just not grade it at all? If the student has done their homework, just consider it passed. And then don't count it at all for the final grade.
And do a final exam on oen and paper and nothing else, and an exam that relies more on thinking than rote memorization of formulas (or print the relevant parts of the course on the exam document itself).
This makes the class really awful. I had professors with schemes like this, maybe calling on someone's name randomly 1-2 times in class to answer a question they'd just asked, just to ensure everyone is scared into paying attention.
Or rather, I do have ideas, but they do require substantial restructuring. E.g.:
1) Strict, in-person (no exceptions), electronics-free exams with written and oral portion.
2) Transition to guided independent learning: students are given written materials and recommended exercises to study. They are shown how to use AI assistants productively. Instead of lectures, there are scheduled discussions, where students can ask questions and listen to additional explanations from the lecturer. They can also seek feedback for completed exercises. However, none of this contributes to grades in any way.
With such an incentive realignment, I think it's possible that higher education could be salvaged. But... apart from general institutional inertia, there is also the political angle standing in the way. Everybody needs a degree in the "developed economy", and the above changes are basically the opposite of the nobody left behind policy that enables half the population to attain one...
1. Make the exam the only thing that determines the grade (or technically, proves that the student learned enough);
2. Give out test assignments students can do, but don't have to;
3. Offer students for the professor to grade their work if they want to get some feedback on how they're doing.
Would that lead to less time wasted by tutors grading AI generated answers? From where I stand, students should be allowed to prepare for exams any way they like, with or without their professor's help. If they managed to learn, they pass.
Of course, this a lot easier for some subjects than others. Subjects like Film Studies relied on exams much less, and assignments much more.
They were disliked by many other students, though, and some people really think university grades should be about things like diligence, compliance, time management, etc., rather than pure subject mastery.
Perhaps (sarcastically) this is the new approach, because this is now the point where they’re forced to struggle and therefore actually learn. An exam every other day, let them ask the questions.
Students will complain to the admins and waste more of your time.
I think this development should make all educational programs reconsider the flip-model: homework at school and video lessons at home.
Homework sucks. Learning should be at your own pace (play/pause/rewind).
In general university lectures are horrible (of course there are exceptions). The lecturures oftentimes are required to teach as a condition of research funds, so for many it's not a priority anyways. On top of that most couldn't care less if their teaching style is didactically sound at all.
Force people to go all in with “impossible” tasks and see how they do with the power of LLMs. Then spend some significant time teaching people _with_ that solution in place what was actually built and what would have been a better architecture? A … software archeology session if you will.
That’s what most of what people do would be anyway - building PoC fast and then figuring out how to scale them?
I know this is absolutely impossible because colleges and universities have headed in completely the opposite direction, but the solution: Literally just expel them. Just like they did when people cheated when I was in college. Don't make it a small punishment so that they turn it into a risk/reward assessment. Colleges are overpopulated anyway; just trim the fat of students who shouldn't be there. Do a follow up oral exam interview about their project and if they can't explain what someone like me spent 30+ hours building and can speak about for 4 hours if asked, then expel them. They can make your coffee at the nearby Starbucks, who cares.
If you don't do this then understand that you are telling your students not using AI that they will need to begin cheating like the others to be able to keep up with the changes. OP mentioned they massively increased the scope of their assignments, which means OP is forcing their students to use AI now.
I wonder what the financial cost of AI will be to teaching? Does teaching get more expensive? More things need to be supervised, more things need to be presentations or in person?
That student presumably performed poorly on the test. Your approach is working as intended here, isn't it?
Also, is the AI-generated exam guide all that different to a human-written guide they could have downloaded a decade ago?
They don't even have to be a significant part of the rubric.
Just ensure the students know what they don't know.
How about cutting out the open book so that the students who don't study and to their homework actually fail? When I was in school many years ago "open book" was only used for special ed students and others who couldn't carry a regular academic workload. Homework was assigned, but our grades were determined solely by testing our mastery of the material. If you could master the material without doing the homework or showing up for class on non-testing days, more power to you. College has become more like kindergarten with all of the hand-holding.
In the classes I took with open note exams where I did well, I barely ended up consulting my notes anyway. The students who didn't study could flip through the textbooks all they wanted, but they lost so much time doing it that they couldn't complete all of the problems on the exam. For the students who studied well, the textbook was way easier to jump around in because they knew it well and knew what they were looking for.
Am exam that is designed to be closed book/closed note where a specific student is allowed some amount of notes as a disability accommodation is just different from one that is designed to be open note/open book from the start. The latter, testing the same material, is generally longer or harder to make better use of the format.
This was too long ago for me to remember the details of but, just making an example here, we might have had a course on homogeneization of elliptic problems using a model PDE with a div(A grad u) term throughout, and then the exam was about a PDE with a curl(A curl u) and some other added terms.
Notes carry the method and theoretical tools but you then still need to apply them to a new problem.
So, if anything, our open book exams were harder and a better test of understanding and mastery. You couldn’t rote your way into a good grade.
If I knew most of my students were using LLMs I would encourage the rest of them to use them too. I would then change the marking criteria such that if they get a single point wrong that's 20% off the entire grade. 5 mistakes gets you a big fat zero.
LLMs are great but they make mistakes. At this point your students would have had to spend so long checking, re-checking and triple checking the LLM's output that they will have accidentally learn what they need to. They will likely need to cross-reference multiple LLMs and at least have read their output which is likely an upgrade on today.
Or some variation of the above, I'd be interested to hear your thoughts.
Reviewing is just not the same as working through a problem yourself. You often can take shortcuts when checking answers for correctness, which generation (with mind or LLM) of the solution can't take. This isn't limited to Math but equally true for programming and many other skills.
The people that do their own work would be at a huge advantage as the LLM(s) can work as reviewers instead of doing the work so there is more chance of catching a mistake before you lose 20%.
On top of that there will be some issues where the LLM is just plain wrong. Being able to work out when that is the case is a hugely useful skill in 2026. It is also something the people who blindly feed their work into an LLM and print its answer without reading it will be unable to discover. Much to their own detriment.
Now this i think applies in lesser degree to the comment I was replying to. With maths, and some comp sci problems, you often can check an answer much faster than actually solving it, eg by using simple substitution for simple math problems. That is a useful skill, but a different one from solving the problem yourself. It employs different types of skills.
That's because we write much smaller quantities of code than we read. Not that LoC is a useful metric, but I'd be surprised if I write more than a few hundred lines a code a day in a legacy codebase. Most of the time is spent planning and deliberating over what to write, and perhaps revising code a few times. Meanwhile, understanding what a snippet of code does can easily spiral into having you read thousands of lines of code across several modules.
But that's for work. writing is a stronger learning tool than reading, and we needed to write tens of thousands of lines of code before we got into the door and started reading more than we wrote. It's still preferable for students to do the same and write as much for their homework as possible.
Explaining that paradigm shift from learning to working isn't trivial, and is often poorly done. But it's needed if we ever want people to frame these tools as productivity boosters, and not shortcuts through tedium. Learning by its nature involves some tedium.
Small nitpick: you teach math, why don't you use an automated grader?
I agree homework should be optional and your college is wrong. Even without AI, for some students it's easy to cheat and maybe for some students it's a waste of time. Maybe you can assign it a very low score, like 5% of the final grade, and/or award the grade based on attempting rather than getting the right answer.
LLM verifies output of another LLM. The circle is complete, and nobody has any doubts that this is just a theatre.
With manual grading, that one student who genuinely tries still has a chance to meaningfully engage.
If the proof isn't trivially translatable to lean, is it actually a good proof?
Perhaps some people do not belong at university? It is just a waste of everyone time, to insist 50% of population studies until 25 years old.
People are choosing to go to university. It might have something to do with having better job options and statistically higher pay…
2. We're also actively giving them less paths in life to begin with. How many high school grads @ 18 will find any job whatsoever without additional school/training, and how else will they afford housing with zero income? The military is the only consistent employer at that age, and the only one who'd provide shelter with the deal.
3. On top of all that, it's in university's best interests to keep 1) and 2) the way they are. Their incentives start at recruitment, and end at enrollment. Past keeping them in college for 4-6 years, they do not care about performance, post-grad job placement, nor societal wellbeing. And part of that is due to a lack of pressure in the recruitment stage to make universities focus on such metrics. Turns out a fancy new gym is a cheaper investment.
Hopefully it doesn’t mean extending high school two years into college, still prepping to make a bet not yet made.
As long as students ask the teacher what they're going to need understanding of maths for they are not prepared adequately.
Reading this has made me understand that the recruitment industry will soon have to adopt the default position that anyone who graduated after mid 2025 is likely misrepresenting the depth of their understanding; people who graduated before that are the equivalent of low-background steel, until our processes catch up.
Maybe there’s still room for me.
The problem is that the student is in charge of the tutor and often tells them to just do their homework for them and not check whether they have learned anything.
LLM providers should formalize a tutoring mode. We should have prompts available to create this mode. This should be the default way students use LLMs. LLMs may not be good enough to do this now, but I think there is already enough raw intelligence and its a matter of training models to be good tutors, to be experts and specific subjects, and developing software systems for tutoring.
In this model there are few lectures. This is more similar to Montessori but with on demand tutoring. The teacher provides objectives and materials for the students and tutors. The tutoring process can be done during class for teachers to ensure it is working well. There can be larger blocks of office hours for students to do their tutoring during.
As AI gets more intelligent, it will also be capable of doing all the grading. This includes oral exams. Teachers need to proctor the exams in person.
Anyone will be able to tutor themselves on any subject. Proctored examinations to certify knowledge do not need to be very expensive. The value of the school environment will not be to transfer knowledge from a teacher but to have peers working on the same thing and teachers and an environment invested in training the next generation.
If you’re forced to attain a pass rate, and can’t watch so many fail, then the certificate becomes worthless in the workplace. Or just offer paid retakes and watch AI make you money.
If you insist on persisting with coursework then a short but heavily-weighted oral part seems to be the only answer, something that will diminish teaching time as a downside, as the article and comments about the flipped classroom touched on.
Students could just use the AI to teach/tutor the material to them instead, but the institutions are clutching the pearls too tightly.
I don't think "learning the material by other means" is the exam being "gamed".
Why is that not desirable? As long as the exam does not allow ai usage during it, i don't see the problem.
The education systems were already hopeless before LLMs and now they got caught with their pants down. Of course some problems were unavoidable, but this overreliance on homework was always a problem in my eyes.
The primary difference now is that the both costs of cheating and ability to verify cheating on the instructor side have decreased dramatically, in tandem. There's also been a kind of mass effect, a cultural change in the attitude of students, where the sheer prevalence of cheating among their peers has made them less likely to feel fear, shame, or guilt over cheating. Students in the past who were caught were likely to confess immediately. Students now are increasingly belligerent and more likely to deny cheating, not just because they believe the professor can't 100% prove it, but also because they believe being caught out for cheating when most of their peers are getting away with it is unfair.
Don't do anything productive for the society but still get to live a decent life on benefits, job security etc combined with addictive social media have ruined most kids will to study.
And the schools are even worse. When I was in school, teachers encouraged learning things; whether for school or outside, with teacher or self help. Everything was fair game and supported. Debate was normal.
Nowadays the teachers in schools are like a cult. Kids are tuned to reject anything from outside; anything that contradicts teacher is a no go zone. Anything that's "too advanced" compared to what teacher is doing is forbidden. Kids are scared of learning by themselves because the teachers punish it. What do you think these kids will do when they grow up?
That strikes me as backward. The higher the expected monetary value of good grades, the more students will prioritise grades over deep learning.
This seems pretty inaccurate to me at just about every turn.
The fire under their asses is for grades and a degree because they've erroneously been told they need one or else they'll be flipping burgers.
It's such a cliche thing to say but cheating at school cheats yourself of the education you could be getting but, since that's not what most people are there for, they cheat. They are there because they've been told that they need a degree to make a decent living and the degree doesn't technically require learning.
I think structuring a course to be taken using a frontier model subscription is outrageous. If you refine the assignments each semester to be only solvable by a frontier model, your students will always be exposed to the actual pricing of said models. It may be 20 USD/month for now (which is already too much to ask in a syllabus in my opinion), but what if prices increase; where do you draw the line? What's next -- make the students pay for cloud credits in DevOps courses?
I agree with the need to increase the weight of oral examinations / interviews with TAs though.
Then there was the written exam at the end of each course, which constituted 100% of the grade. The only reason that got abolished was because it made too many students fail, and the board hates that. So that became a race to the bottom, and apparently, we can still go lower.
It's reaching levels where most uni diplomas don't matter any longer. The only thing that will then matter is your network. That's the end of social mobility.
The article gives examples of LLMs being wrong. An AI grading of an AI submission would presumably give a good mark, to a wrong answer.
I think I understand what you mean though, and I kind of agree. I just don't like the use of the word chaff, like something you would discard permanently.
I'm pretty sure I was "chaff" at some point (maybe multiple) in my life growing up.
"This is math, you can find all solutions to the course problems in old material since I've been teaching here for 20 years, but know this you're cheating yourself and at the end of the semester you're taking a three hour test on pen and paper so good luck"
I see zero reason why AI should impact teaching. You can grade exercises as information to the students as to how well they're doing, but simply have an exam at the end of the course. Needlessly to say like 60-70% flunked through Analysis I and Linear Algebra. But they're adults and they're there voluntarily, so why is this an issue
My hope is that this is a wake up call to teachers / professors to find a better way to evaluate students. To determine who has actually come away with a real understanding, vs those who were running around copying the assignments off of their peers anyway. These same people are now just skipping the peers and going straight to the all knowing oracle. It's the same as it ever was, degree mills.
The example questions given in the article do not assess wrote learning.
> It's been distilled to the point that the lowest of the lowest common denominators can pass.
Have pass rates been increasing? Unless I missed it, the article doesn't say so.
> forced to learn things they will never use
A computer science degree is not a coding academy, it's a basic grounding in an area of study. A computer science graduate should have some understanding of theoretical computer science and of computer architecture, say, even if they're unlikely to apply these topics directly in their careers.
> My hope is that this is a wake up call to teachers / professors to find a better way to evaluate students. To determine who has actually come away with a real understanding, vs those who were running around copying the assignments off of their peers anyway.
Academics are already aware of the importance of fair assessment, but it isn't easy, especially with LLMs in the mix.
> These same people are now just skipping the peers and going straight to the all knowing oracle. It's the same as it ever was, degree mills.
Students that cheat, and degree mills, are two different things.
I frequently see this in remote interviews for senior developers. These are conversational interviews, not coding tests, but some candidates seem incapable of having a conversation without AI.
I can't imagine how these candidates would be able to meaningfully contribute to a whiteboard design discussion with real people.
For a previous client, we resorted to a simple in-person pen-and-paper screening questionnaire with 5 very basic questions. Many candidates declined to attend, and most of those that attended failed dismally.
- some skills only develop through the actual writing (similar to how you can't learn to code just by watching YouTube videos)
- in-class assignments are not always scalable (and students end up using AI anyway)
- students aren't motivated enough to participate actively
I think looking at the writing process is a more realistic approach, so I built Turingo (https://www.turingo.net). It replays how a Google Doc was written, highlights unusual edits (large pastes) and tracks how that text changed afterwards. Surprisingly, several professors reported that students often underestimate how much LLM-generated text is left in their final drafts, so just seeing the replay changes the conversation.
It's not meant to be a verdict machine like the existing AI detectors (another big problem in education), more of a starting point for the kind of conversations the author's TAs are having. In fact, we're also beta testing a feature that suggests a few questions for the professor to ask based on the replay.
And yes, it's not perfect and autotypers do exist, but they're dumber than you'd expect and nowhere close to mimicking a real human (I hope to address that soon).
Interestingly, this applies not just for learning, but also to a great extent in production/work contexts too.
Isn't the answer the 'flipped classroom' - students are assigned reading/studying to do before a class and the class becomes more interactive, answering questions/discussing topics/solving problems etc depending on subject. Of course, this is more expensive than a hall with 400 people listening to a lecture with a couple of tests/essays.
https://fltmag.com/the-flipped-classroom/
https://en.wikipedia.org/wiki/Flipped_classroom
I've heard of places where the minimum points for an assignment is 50% even if you never turn it in which is insanity.
My main point is: flipped classroom works great in some contexts but is no panacea and requires setting clear expectations.
Aside: as English is not my native language: "you fail them" can mean both "you give them a failing grade" and "you fail to provide the support they need", right?
Yes, but as written the latter would not be a reasonable parse. For that sense: “you have failed them”, sure; “you are failing them”, sure; even “in acting so, you fail them”; but “and then you fail them” as an entire sentence would be very weird. The emphasis I placed on “fail” also doesn’t feel particularly compatible with that sense of the word.
I also felt having multiple explanations (in a regular class where you could also read the book in advance) from regular lecturers helped me in gaining understanding because having something explained in multiple ways made sure there would be at least one way, or a combination, which made it 'click'.
I always liked the regular lectures where you would be required to do some pre class reading and a smaller set of exercises the most, as you would be engaged in the class itself and would be getting repetition with different explanations.
Flipped classrooms have incredibly weak causal evidence. However they are popular in spite of this - perhaps because of this if you're cynical - as the educational research field is pretty questionable in its standards and the profit motive is a strong thing lurking in the background. Wikipedia has a big warning on the article for a reason, but even then it has serious issues.
"Active learning" is a thing (insofar as it has meaning rather than be a useless buzzword catch-all) that does work. But the "flip" specifically does not, or rather, sometimes a mix of extra effort on behalf of the teacher and other loosely related effects occasionally can make a modest impact that makes up for its failings.
Most of the people pushing flipped classrooms are the same people who think that students learn best when they "discover" concepts on their own, maybe with a bit of guidance. This is the other, more serious, lie in education right now, akin to the whole idea that you didn't need to teach phonics to kids when they learn to read.
It’s a virtuous circle though, as many universities are there for the money, so it’s a beautiful exchange.
But you're not doing that, you've just given up and allowed any use!
> How to use AI coding agents effectively is not a learning goal
Exactly
> violate evidence-based best pedagogy practices, and I made them anyway.
Of course you have, when have educators cared much about best practices
> We have since built infrastructure to autograde code and written reports of homework with LLMs (institutionally approved LLMs)
Further reducing the point of your own existence
> Debriefing after such a failure is a good learning opportunity to talk about automation bias and different forms of human oversight (and how this is really hard in practice).
It would be if there were any learning, but since the incentives are for more AI use and less learning, there will be no application of the supposed learning, so you'll continue to have those 80% fails
Why do you say so?
This is not at all my experience, as a recent teacher it strikes me how much the teaching methods are supported by science.
* Alternate between theory and practice in the course of a single 2-hours slot. Theory is necessary, but the attention span of people is very limited. * Grade either a homework assignment in the form of a real project with specifications. You give the assignment half-semester, so that students who start early can ask questions. Otherwise, grade a in-session exam, with every resource available, including the course and the whole Internet. However, you don't assess the knowledge, but if the student is able to apply their knowledge to the different tasks at hand.
These approaches are now completely moot in the age of AI.
I was surprised the first time a colleague asked me to do an oral exam. I thought it was a burden on people with poor social interaction skills. It went surprisingly well: you could see in a couple of minutes if the student had integrated the concepts of the course, even if they were shy. Now I wonder if it's the way to go. There's one caveat, though: it doesn't scale.
There’s no way I’d have learnt everything I know now about software in today’s environment. So much of software development - debugging techniques, architectural decisions, structuring data etc is learned through trial and error. As the models get better is becoming easier and easier to just not look too closely at their output. I suspect it’s human nature.
2 hours is a lot! In my universities, classes were either 50 minutes, or 70 minutes. I could handle 50 minute ones just fine, but would often have trouble with challenging courses that had 70 minutes. Particularly for stuff like math, I hadn't properly digested the material in the first half, and now he's already proving theorems with that material in the 2nd half. I can follow the logic, but the ideal is when you can digest the prior material before being exposed to theorems relying on them.
Even when alternating theory/application like you say, the amount of attention I'd give to the application in the 70 minute class was less than the 50 minute one. I could follow, but I wouldn't think critically.
Usually you know after a few minutes if it's going to be a fail, and then otherwise you only need to figure out where on the passing scale the student ends up.
They've shown that it is entirely possible to scale a system like that, as long as the society sincerely values the role of an educator.
In my very limited experience (I have German major degrees in Informatik and Philosophy, I was a TA for Logic in Philosophy) the more advanced the topic the thinner the attendance anyway.
Finally, if a student feels their result is unjustified, reevaluate the recorded session.
The end goal of education is to help you understand how things work, with problem-solving being a means of developing and applying that understanding. LLMs shortcut the former and turn the latter into a pointless arms race that we cannot win.
I loved the example of increasing the complexity of learning projects, but I still feel like one question is missing: are people really trying to learn the same thing we were trying to teach them in the first place?
This is how we do PRs now.
Why would AI undermine quizzes? I took an intro to Japanese class where we would have a quiz at the beginning of each class which I quite enjoyed. I can’t see how AI would change the utility of that.
It was also interesting to see that everyone is now forced to use LLMs to do the assignments. I wonder if students are using any kind of subsidised access to LLMs here.
This gives me/us a lot of angles to understand what progress has been made throughout the semester as well as to consider and react to individual differences between students. In my book, it’s also resistant to faking it convincingly throughout all modalities. However, scaling that approach to big entry level courses would require more money for teaching than any research group in Europe I’ve ever seen has available.
If someone manages to learn without doing the homework, or finds some way to learn better using an LLM, fine. The exam will show that.
— watches, calculators, discrete technological devices. I've seen kids encode text as patterns of dashes on their desks, sharing the cipher solution with their desk mates in other periods.
The problem isn't skirting the effort. The problem is systemic. Why does the system reward taking shortcuts? I think it's because most areas are focused on short horizon tasks, min-maxing effort:output. It's not a problem with those technical fields. It's a problem with the current crop of the C-Suite across the entire occupational plane. They're short sighted.
Once you have a wishing well - i.e. you own an eduction in baking or dentistry or law or math or whatever - you can ask it anything and it will grant you your wish. If others want something from your well, they have to come and ask you. What you give them has your name attached to it. So, your reputation rides on the quality of what your wishing well produces.
Prior to AI, universities chose which wishing well they would specialize in building and offered to share that knowledge with students.
In the age of AI, the creation of personal wishing wells has become trivial but universities still think they are in the business of teaching people how to create wishing wells.
Ultimately, the only truth left is for owners of the wishing well to start offering the product of their wishing well to the market and attach their name to it. If there is a market for their product, great. If not, go make another wishing well.
Education used to be pets. Now, it is livestock. Universities are not willing to accept this shift in mindset.
Lest anyone think this is a brilliant idea: an LLM would likely be much better at doing the ~~RLHF~~ RLMF, so you haven't actually gained anything. But I feel like there may still be the kernel of a good idea somewhere in here.
Perhaps your assignment is to iteratively train the AI, which is pre-prompted to seek out every reasonably possible failure mode when it generates solutions. So the learning mechanism is to train an adversarial AI on the subject matter as a way of learning it yourself. This does not solve the problem of cheating with an LLM, it's an alternative pedagogical approach.
The only people really having a crisis here are teachers who want to feel relevant.
Similarly, I did well in interviews where I could use the internet, but I didn't do well in interviews where I couldn't use the internet.
Everyone has a testing format they're good at. I'm curious about how this kind of thing is evaluated, and whether there are any papers related to this.
I recently did a course on teaching bachelor-level education, and had to study a couple of books. They all seemed to quote their sources and I got the impression that they're rooted in science.
I think society is moving back toward nature at rapid pace. More adversarial in general, and advantage taking behavior is the way to survive if you don't have a solid community. Society will be culled to a much smaller percentage of people than are present today, and it probably won't be pretty. Smaller communities instead of larger communities, with self-sufficient systems in place and a lot of warfare...
Don't see how AI overlords controlling everything will work out. If no one can buy the food from the large producers, then they will dissolve too. Supply chain will fail at some point in the next couple years maybe? Diesel is already outpacing wages for truckers.
The people with all the money are positioning themselves to reap the assets from others to give themselves as much leverage as possible.
At this point I don't really see how it could go any other way.
I don't know. I'm trying to figure out what's going to happen, but it doesn't really seem like anything good is going to happen because the trend is wanton disregard for morality in exchange for stealing wealth.
Makes me really sad. I'm also on fixed income so I worry quite a lot about my future.
AI can talk the talk, but it can't walk the walk.
A simple "explain why this is the best solution" will make it shit the bed 10 out 10 times.
Clearly teaching got a lot better, since we now have a new teaching tool at our use, the LLMs. It's evaluating that got harder, because students can use the same tool to cheat in traditional forms of evaluation.
If "teaching the students" means the same as "evaluating the students" to you, maybe you shouldn't teach in the first place.
Even devoid of all the economic and social pressures that grades impose students would compromise their learning experience. Ever regretted looking up the solution to a puzzle in a video game or sneaked a peak at the crossword puzzle solutions?
One of the hardest part of teaching is keeping students from self-sabotage. Evaluation helps with that.
Beyond that evaluation is of course also useful just as feedback.
It’s like copy-paste Wikipedia presentations 2 decades ago. People used to learn things doing research for a project. Then suddenly it became copying off Wikipedia and people not even pre-reading what they copied into PPT.
Only because you're obsessed with being able to assign grades and fail cheaters.
No obsession required.
The situation reminds me of what my teachers used to tell me. "Learning how to look up things in a book will be a skill you'll need all your live and you won't carry a calculator with you at all times when you are grown up." Both assumptions were proven wrong before I was even out of school.
If the only remaining value of education is "to become a well rounded person" then why waste your time on it? It used to be you went to university because that was the only place to get knowledge from. Which is also why it was dominated by the rich and powerful. These days university feels a lot like paper money. Everyone just pretends it has a real value and because everyone agrees, that's where the value comes from.
The ones who don't want to learn will cheat themselves of an education. This is their own problem and we shouldn't punish the actual students to catch these others.