36 comments

[ 3.3 ms ] story [ 105 ms ] thread
So the business school awarded money to develop a morally ambiguous (but likely profitable) "study aide" app, and the school ethics board said that creating and using it violated their cheating policies.

Sounds like the most realistic lesson that business school could have taught?

Yeah, I though the same. These students will probably more careful in real life.
Yeah, that'll teach them for teaching the wrong way
There don't seem to be any indications that it's morally ambiguous. The university apparently got extremely upset that it connected to Canvas (the online assignment platform), but from the lawsuit filing in another comment, it only did that to pull (not submit) documents to turn into flashcards or whatever.
Did it pull college course material?

That could be a copyright violation.

What's used as the backend AI? OpenAI? If then it's more free data for the company that doesn't value people's rights.

It feels like there are more specific accusations of misconduct against the student/app that are missing from the article...

The university claimed it violated their ethics policies, but for what reason? Further the article states there is no evidence directly linking the app to cheating?

It's a bit strange the student in question submitted letters of apologies... For what was he apologizing for? Feels like there are some details missing.

I dunno, Universities can often behave like soviets. I was pushed through a university kangaroo court at Northeastern where they tried to (falsely) pin a charge of destruction of the university wifi network on me. They were threatening suspension and expulsion, I had to go to numerous hearings attended by high ranking university officials, and was definitely told that writing a letter of apology would incline the university to lighten my punishment.

I fought my way tooth and nail out of the situation and was exonerated, but I can imagine a hundred reasons why any given student in a similar situation would capitulate.

> I had to go to numerous hearings attended by high ranking university officials, and was definitely told that writing a letter of apology would incline the university to lighten my punishment.

This reminds me of the police telling people that writing out a confession will get them a lighter punishment.

I know a couple of people who were dubiously accused of network related shenanigans at college, and yeah - you are heavily pushed into confessing to something, there’s no due process, fuck all burden of proof, etc.

Truth doesn’t matter in university disciplinary matters - only finding someone (not the university) to blame.

(comment deleted)
I remember watching a news magazine (may have been 60 Minutes -long time ago), that talked about how universities handle sexual assault allegations.

They have their own police force, and the segment was about how they would often keep regular cops frozen out of the investigations, and just wanted to "hush things up," often at the expense of the victims.

It’s the same with theft, normal assault, fraud, normal criminal behaviour, etc.

Universities often want to keep it “in house” and deal with it through internal disciplinary actions - as opposed to bringing in the law enforcement and public scrutiny.

This exists outside the US too - and doesn’t require a campus police force.

In the university a friend went to, such matters (theft from accommodations, etc) were dealt with through the deans office, and calling the cops was heavily discouraged by the university.

This system has pros and cons, but when it’s to do with more serious matters (actual assault, sexual assault, etc) the real authorities should be brought in instead of some bizarre victim silencing and due process lacking internal system.

Sometimes, universities just disseminate this information to their students, too.

Somebody hit to our vehicle (owned by a government institution) and tried to ran away into their university, but failed. We had to call the police (since that's a government institution vehicle), and university security guards came and told the driver "if something like this happens again, just try harder to enter university premises, we can help you".

We politely told the guards that this is a government owned van, and we need to call the police, anyway. They frowned very visibly to this fact.

The driver was also very sad about not being able to outrun us.

This was not in US.

> This was not in US.

If it was, I’d have just let them get away, and called the cops anyway.

They would be charged with leaving the scene of an accident (hit and run). The campus rent-a-cops would not be able to do a thing.

Same would have happened if the driver have managed to run away, but lucky for them, they failed. Otherwise, it'd be much bigger fine for them.
I was a student on one of those university boards. I highly recommend hiring a lawyer at the first hint of an issue.

My university wanted people to accept "responsibility" which students took as responsibility for their actions, not for what they were being pinned for.

Of course the students are going to take responsibility for their actions thinking that will help, and then the University does whatever with its mind already made up.

One student had trace amounts of drugs on some paraphernalia. Student didn't show up to their board hearing, so got the default suspension. Could have stayed and brought a lawyer for less than what the semester cost (instead of receiving Incompletes/Fails) and won. Oh well.

(comment deleted)
Yes this happened to me too, but the rounds of applause I get on Internet forums makes it all worth it!
My guess - the app's users uploading university course material to a third party is likely a violation of school rules. Emory IT disabled the button to generate Canvas access tokens, and the 'workaround' the kid developed in response caused the IT folks to escalate.

An LLM trained on course material could in theory be used to generate homework answers, so they wanted to shut it down before it actually happens. (Hell, the demo question on their website[1] sounds taken straight from a problem set) It would be as if the answer key to the homework leaked, except its for every homework in every course. The punishments are to discourage others from building clones. Though I suspect this is going to cause a Streisand effect.

[1] https://www.eightball.ai/

This is somewhat outdated. Friday the judge ruled the suit can not continue anonymously. https://storage.courtlistener.com/recap/gov.uscourts.gand.32...

> The Court does not intend for this determination to suggest that all information in this case will or should necessarily be made public. Indeed, there are other ways to protect specific information in litigation. See, e.g., Fed. R. Civ. P. 26(c) (allowing for protective orders for “protect[ing] a party or person from annoyance, embarrassment, oppression, or undue burden or expense” upon a showing of “good cause”). The fact that these other means exist and are—in the Court’s view—better suited to address Plaintiff’s instant concerns also indicates that Plaintiff proceeding anonymously is not required to achieve the protection he seeks. Thus, the Court makes clear that while it will not allow Plaintiff to proceed anonymously

And then indeed he refilled and so his name is now known https://www.documentcloud.org/documents/24676044-16

Reading the lawsuit it seems another student, who apparently coded the thing connected Canvas to Eightball and that's what this whole thing is about. The lawsuit alleges Plaintiff had no idea this happened.

Tangential but, why shouldn't universities be afraid? The value of higher education for many subjects is existentially threatened by AI. In 10-20 years, it's not unrealistic that the only justifiable subset of subjects will be those whose practitioners cannot be replaced by AI.
The pedagogy needs to change. Some coursework should embrace new tools and focus on how to be effective while using them. And some changes should take place to the balance of exams, writing assignments, projects, presentations to reflect an honest evaluation of the student.
People said the same thing about the internet and the "highway of information" and see how that turned out.

The value of a College Degree isn't the content alone, but the mental models that are taught. That's why it's used as a credentialing process.

As someone who earns more than my degree-holding parents through self-taught skills from the internet, AI learning seems to have a bright future.
And statistically speaking, how many degree holders versus non-degree holders end up with 6 figure jobs?

I'm not denying that you can't self learn, but most people don't have the discipline and drive needed to reach that place. Nor will the self learner pass the hiring bar for a junior role in 2024 anymore.

Non-traditional college education (eg. low cost online education, expanded community college access, asynchronous learning, etc) is a happy medium and was in fact the norm for college education in the US throughout much of the 20th century (the normalization of the residential dorm model is a newish innovation - most students were historically commuter students from the early 20th century with CUNY all the way to the CSUs and UC extensions of the late 20th century).

Programs like ASU or OSU's online BSCS are a better middle ground.

If you are sick, do you visit a doctor or the guy who taught himself through the internet?

If you cross a bridge do you have the same kind of trust in its stability if it's constructed and inspected by engineers with a college degree or some guys who played bridge builder?

both of those you can't, because of government regulations.

But (at least in the US) you go to your smart friend who taught himself through the Internet unless you're lucky enough to have actual doctor friend before you step foot in a doctor's office for anything weird because otherwise it's a waste of a copay. If I go to the doctor who went to doctor school and say I have this and that, they're gonna run some random tests and then you have to go back in a second time to run more tests. If you go into the office advocating for specific tests you want run, with scientific papers as reasons for wanting them run, then you'll get the tests you need the first time around, with informative results (positive or negative or wherever you are on the scale) in fewer visits than if you didn't ask your smart friend. If you go into a specialist doctor's office without having a list of questions to ask, for your problem, you're not getting good value for your time/money. Covid exposed a number of shortcomings in the system, and if your malidy is at all weird, you're screwed.

> you go to your smart friend who taught himself through the Internet unless you're lucky enough to have actual doctor friend before you step foot in a doctor's office for anything weird because otherwise it's a waste of a copay

This is by definition pennywise pound foolish, and this is a very specific example, so I think it's something you do.

I'm going to assume you are a decently paid white collar worker (earning $70k+ a year) and tell you that you should not think this way.

I'm going to assume you don't have any friends with ME/CFS/fibromyalgia/Long Covid or anything "interesting" and tell you that you're lucky to not have had to advocate hard for what people would consider decent level of care in the system yet.

(only 70?)

THe value of a college degree is signaling.

Your ability to graduate college signals to future employers that you have some combination of intelligence and work ethics. The more prestigious the institution, the stronger this signal is.

At the top end of the scale, it's not even about course difficulty but admission requirements. A lot more people would be able to graduate from Harvard if given a place than can actually get into Harvard, legacy admissions who are still able to graduate being the perfect proof of this. There's research that shows that Harvard courses aren't that much more difficult than those at other, far less prestigious universities, but the admission requirements are far stricter.

This phenomenon is a self-perpetuating feedback loop. If your average college graduate is slightly smarter than a non-graduate, employers will prefer graduates over non-graduates, as it's an easy thing to check. People notice how much of an advantage the college graduates have, so in the next generation, even more people choose college over other paths. This makes the gap between graduates and non-graduates even wider, which makes employers favor them more, which makes life harder for non-graduates, which causes even more people to go to college, and so it goes.

At the top end, other things start to matter, it's advantageous for smart kids to mix and network with rich kids and vice versa, hence legacy admissions, but this is what matters for most graduates.

> THe value of a college degree is signaling.

It really isn't. If so, why the hell do UC Berkeley, SJSU, and UIUC CS grads outcompete Dartmouth grads in YC, Entrepreneurship in general, or the tech industry?

Yes Harvard-type programs are exclusive, but they do not open that many more doors in our industry than a CalPoly or UW or Cal undergrad would have. And the college advantage goes away very fast with 4-5 years of work experience.

I don't need code monkeys. Learning to code isn't a big deal. If that's all I needed, I'll just contract to Infosys or EPAM. What I and other employers need are CRITICAL THINKERS - people who can think on their feet about both technical and business problems.

This requires building that critical thinking muscle. Some people learn that out in the wild, but a lot don't, and college is the first place they learn to build that muscle. College is also one of the first places you learn that you can't coast - for every grade inflating Harvard, there are 10 grade deflating UCs or public universities.

We shouldn't put barriers for hard working people who have self taught themselves the ins and outs of the industry, but we shouldn't conversely automatically shit on the entire concept of tertiary education.

Remember that AI was supposed to be solved by McCarty, Minsky and Shannon in the hot summer of 1956.
(comment deleted)