Why This Teacher Has Adopted an Open ChatGPT Policy (npr.org) 113
An anonymous reader quotes a report from NPR: Ethan Mollick has a message for the humans and the machines: can't we all just get along? After all, we are now officially in an A.I. world and we're going to have to share it, reasons the associate professor at the University of Pennsylvania's prestigious Wharton School. "This was a sudden change, right? There is a lot of good stuff that we are going to have to do differently, but I think we could solve the problems of how we teach people to write in a world with ChatGPT," Mollick told NPR. [...] This year, Mollick is not only allowing his students to use ChatGPT, they are required to. And he has formally adopted an A.I. policy into his syllabus for the first time.
He teaches classes in entrepreneurship and innovation, and said the early indications were the move was going great. "The truth is, I probably couldn't have stopped them even if I didn't require it," Mollick said. This week he ran a session where students were asked to come up with ideas for their class project. Almost everyone had ChatGPT running and were asking it to generate projects, and then they interrogated the bot's ideas with further prompts. "And the ideas so far are great, partially as a result of that set of interactions," Mollick said. He readily admits he alternates between enthusiasm and anxiety about how artificial intelligence can change assessments in the classroom, but he believes educators need to move with the times. "We taught people how to do math in a world with calculators," he said. Now the challenge is for educators to teach students how the world has changed again, and how they can adapt to that.
Mollick's new policy states that using A.I. is an "emerging skill"; that it can be wrong and students should check its results against other sources; and that they will be responsible for any errors or omissions provided by the tool. And, perhaps most importantly, students need to acknowledge when and how they have used it. "Failure to do so is in violation of academic honesty policies," the policy reads. [...] "I think everybody is cheating ... I mean, it's happening. So what I'm asking students to do is just be honest with me," he said. "Tell me what they use ChatGPT for, tell me what they used as prompts to get it to do what they want, and that's all I'm asking from them. We're in a world where this is happening, but now it's just going to be at an even grander scale." "I don't think human nature changes as a result of ChatGPT. I think capability did."
He teaches classes in entrepreneurship and innovation, and said the early indications were the move was going great. "The truth is, I probably couldn't have stopped them even if I didn't require it," Mollick said. This week he ran a session where students were asked to come up with ideas for their class project. Almost everyone had ChatGPT running and were asking it to generate projects, and then they interrogated the bot's ideas with further prompts. "And the ideas so far are great, partially as a result of that set of interactions," Mollick said. He readily admits he alternates between enthusiasm and anxiety about how artificial intelligence can change assessments in the classroom, but he believes educators need to move with the times. "We taught people how to do math in a world with calculators," he said. Now the challenge is for educators to teach students how the world has changed again, and how they can adapt to that.
Mollick's new policy states that using A.I. is an "emerging skill"; that it can be wrong and students should check its results against other sources; and that they will be responsible for any errors or omissions provided by the tool. And, perhaps most importantly, students need to acknowledge when and how they have used it. "Failure to do so is in violation of academic honesty policies," the policy reads. [...] "I think everybody is cheating ... I mean, it's happening. So what I'm asking students to do is just be honest with me," he said. "Tell me what they use ChatGPT for, tell me what they used as prompts to get it to do what they want, and that's all I'm asking from them. We're in a world where this is happening, but now it's just going to be at an even grander scale." "I don't think human nature changes as a result of ChatGPT. I think capability did."
Humans will never have to think again (Score:3, Funny)
Re:Humans will never have to think again (Score:5, Insightful)
When tutors read students' submissions & see perfectly formed texts that are mostly factually correct but at the wrong level of analysis, irrelevant to the task/problem/question, etc.. Lazy, hurried, poorly informed prompts will give generic & irrelevant responses. All tutors have to do is ask a couple of questions directly to the student to see if they actually wrote it.
Re: (Score:2)
Re: (Score:2)
My experience as well. Arguing with it and adding details can make large differences to the usefulness of the output. Not to mention needing the knowledge to tell when it has gone off the rails or at least where to go to check its output.
Re: (Score:1)
Well, yes. The reason this is all so controversial, is because our educational standards have sunk so low that dismal results are often considered good enough. This doesn't mean ChatGPT is smart. It means too many of our students are being passed along through the educational system without actually learning to any meaningful extent. If they can parrot back a few stock answers from a "World Almanac and Book of Facts" and/or Wikipedia, t
Re:Humans will never have to think again (Score:5, Interesting)
I think the point is for ChatGPT to come up with ideas which the students then have to critique and think about, not just parrot.
Good starting point for critical thinking.
Re: (Score:3)
30 years ago, it was important to know how manipulate variables and memorize rule for calculus. Engage in row operations for linear algebra. Manipulate large data sets for statistics.
Now one has to ask questions, set up equations, and interpret those result. One still has to know the logic so one can justify the result as so[[valid and sound. Medical research demonstrates what happens when we reach conclusions without context or dee
Re: (Score:2)
Humans will never have to think again
Thinking is exactly what ChatGPT doesn't do. At all. It cannot draw logical inferences, can't even do simple arithmetic. It has a lot of encoded knowledge and is good at generating well-formed, reasonably-sensible text. It occasionally generates good ideas, but it has no way to distinguish good ideas from bad ones, good arguments from bad ones, or true facts from falsehoods -- no ability whatsoever to evaluate critically.
Re: (Score:2)
More specifically, I'd say deduction pad the skill it lacks.
If you use the Google extension , it can certainly summarise those results -- a conceptual thought process.
ChatGPT can solve problems (Score:2)
I asked it what resister I needed to reduce 12 volts to 10 give a 1 amp load.
ChatGPT produced an impressive summary of the physics, the formulas, applied them, and confidently provided the answer: 0.2 Ohms.
(I hope slashdot readers can still see the error in this.)
It misinterpreted my question, misused the formula. But I was still very impressed.
However, as this becomes more popular, I foresee bridges falling down! I doubt many young engineers know their times table today, in future they will not know wha
Re: (Score:2)
I asked it what resister I needed to reduce 12 volts to 10 give a 1 amp load.
resister
noun
: one that resists
especially : one who actively opposes the policies of a government
Maybe it gave you the right answer?
Re: (Score:2)
How many humans do?
I still remember one of my surprises with ChatGPT. It generated some sample code with a comical mistake that nobody awake could have made.
So I tried something that absolutely should not have worked. I asked "Does this code have a bug?" and pasted in what it had supplied.
It reported the comical mistake, accurately.
It is a strange alien thing and I'm still working out what it can and can't do.
Re: (Score:2)
Math is not sums unless you work in a bar and even then the arithmetic is usually done on a till these days. A lot of "math" now will no longer require you to memorise formula with AI available - but for example you will still have to understand the difference between different statistical tools even if the exact formula escapes your memory.
Next logic step (Score:3, Insightful)
Re: Next logic step (Score:5, Insightful)
I tell students who are about to graduate that colleges aren't gonna be relevant in a few years, and instead people will learn at their own pace and from any instructor that they like. Colleges and universities should just become test centers that award certifications.
I only see a case for vocational schools that teach hands-on skills going forward.
Re: Next logic step (Score:5, Insightful)
I've been a teacher 20 years, but began 30 years ago and with an interest in cognition and psychology of education.
Classic education has been forgotten and is not obsolete. Reading PowerPoint slides is modern but it was never progress. Oral exams are so old that nobody in the USA has even experienced them and they cringe at the thought. Actual interaction with the educator / professor is all but dead and personal study is low-- it's actually harder for all of us to READ because our brains have been trained to skim and filter so much linear reading is uncomfortable (even heavy readers sometimes notice this.) Attention spans and addictions are way up and now I see students ...even retired people... who are constantly "busy" with "so much to do" when they actually have nothing going on... just tech addictions like a massive Netflix queue they feel looming above them getting bigger not smaller. We are the most entertained and distracted humans in history and the last 2 generations have a shorter attention span than a goldfish (actually an experiment.) I see FEW students with actual real hobbies anymore; consumption "hobbies" are the majority of hobbies.
Yes, I shifted topic into major cultural problems because they are far more damaging than the devolution of modern education. The answer is not to adapt but to enforce and fight back; you can't win allowing kids to use calculators before they learned how to add. You ban them initially and restrict harmful things until they can be handled, if ever in that environment. Smartphones lower YOUR IQ simply having them on you, they should be banned from all classrooms like guns are. I've done online education during COVID with open access to everything and a pile of hands-on work to complete and me as the tutor and evaluator -- it takes a lot to keep the work from being pasted from online answers and to verify it has been completed. (most of it was instant feedback so grading is reduced to verification.) You NEED deliberate practice and repetition. Most people especially today require somebody to force them to do this as well as some guide to provide a path to mastery (otherwise you become a niche expert nobody wants and lack fundamentals.)
The master apprentice model which predates schools themselves is alive in niche areas and a vastly superior approach to adapt from if you want to reboot the system. It certainly does align well with positive use of technology - but it can work classically as well if you look at trade unions who have classes to take and textbooks spread out over the career's progression instead of cramming it ALL in BEFORE they even begin in the trade. Some people NEED to be forced into a schedule; I have been so open ended, unconventional, and tried so many things-- I have a significant number of people who can not function without a task master and it is growing because they feel so "busy" (but are not...priorities are foobar.)
Hands on education can happen for any topic in any classroom. It no longer does. It didn't always happen but far more - especially in college when it only consisted of the elite - now a system decided for elite learners is being applied to the masses and with tech that does far more harm than good. Some people like myself are self-taught and do better on their own... although they may not good test takers; way too often, I've run circles around certified people... I could always use a mentor or helper or coach to make my self-learning more productive. Self-assigned projects are great but rarely are comprehensive.
It used to be correspondence school was a joke of an education...but better than nothing. Now we put "cyber" in front of that and it's all hip and great. but do we know the difference when we have grown up in a degraded experience? people don't know online sucks because in person has sucked their whole life. Multiple guess exams are bad maybe toxic. Memorization that is not important but we assume it works because historically elite student recall reflected a deeper
Re: (Score:1)
I must say I pity your students over the past 20 years if your teaching style is anything like your /. posting style.
Everything in your comment could have been communicated just as effectively with half the words and better organization.
Your post is ironic om that it is critical of the fact that, supposedly, our "brains have been trained to skim and filter" - yet your comment encourages, indeed almost requires, people to do just that in hopes of finding useful content and a salient point somewhere in the se
Re: Next logic step (Score:2)
Still, I found it inspiring
Re: Next logic step (Score:1)
Re: (Score:2)
I can't wait to see how close chatGPT is to those students... My theory is they will be surprisingly close. We know the AI doesn't understand anything but repeats the highly complex non-linear approximations it searches for in the problem space that "pleases" the testing. The passive student can just be aping patterns they picked up with minimal effort and the more powerful tools the less effort required; therefore, the less relevant those simpler patterns will be to actual understanding.
Maybe AI mirrors ou
Re: (Score:2)
This is a /. rant, not an article, paper, or a proof read SA.
Re: (Score:2)
Please anonymous coward, come show me a half size version with the same ideas.
Re: (Score:2)
On evaluation, that's what I criticize about classical education. I have come across some very bright students, some of them even helped me teach the class,
Re: (Score:2)
I actually have shown that TED Talk you linked to in my class since it was posted online. I'd say it gets about half the class thinking; I can't ban distractions or more would grasp it. It almost never is a matter of IQ but focus and willingness to activate their brains -- when their habit is to idle their brain just high enough to positively stimulate their System 1 (autopilot.) There are great TED videos on irrational behavior (and a couple books) on how we don't really consciously make decisions like
Re: (Score:2)
I think you may enjoy "Punished by Rewards" by Kohn.
Re: (Score:2)
You should open a book some time. There has been quite a bit of solid research and a ton of innovation over the last 20 years that you seem to have missed completely.
It used to be correspondence school was a joke of an education...but better than nothing. Now we put "cyber" in front of that and it's all hip and great.
If you think online learning is just the old correspondence model but 'on the internet', you're not qualified to have an opinion. Try taking a few online classes. Odds are good that you'll engage more with the instructor and your classmates than you ever did in a traditional classroom.
A lot of teachers are using hybrid online/classroom mod
Re: (Score:1)
Interesting read, thank you. Agree with everything except for this:
The answer is not to adapt but to enforce and fight back
I don't think that's possible. The technology is here and is easy to use, I don't think you can stuff this genie back into the bottle now. No more than Victorian age factory workers throwing their shoes into the machines in order to break them could stop the Industrial Revolution. For better of for worse, this thing is here now and we'll have to adapt, if only to try and minimize the damage.
Re: Next logic step (Score:3)
Re: (Score:2)
As I mentioned, there is a lot of garbage in formal education, like having to be stuck with one instructor and one specific method of evaluation (usually multiple-choice because you can let a computer grade for you, and I've always hated multiple-choice as a student and a teacher). You are al
Re: (Score:2)
they can't see the graded final exam paper unless they make a complaint and risk the teacher nitpicking even more with their grading.
You kind of answered your own question earlier:
Colleges and universities should just become test centers that award certifications.
Test centers don't want their exam questions to leak. The accuracy of their assessments is very important. Teaching the material isn't their goal.
Today's colleges try to be something in between test centers and a purely educational institution, and those 2 have conflicting goals.
Re: (Score:3, Insightful)
Are you kidding me? The idea that we could replace dedicated, compassionate human teachers with a machine is ridiculous and insulting. AI may be able to provide information, but it lacks the ability to truly understand and connect with students. It cannot provide the personalized instruction, emotional support, and critical thinking skills that are necessary for a well-rounded education. Suggesting that we replace teachers with AI shows a complete lack of understanding of the important role that teachers pl
Re:Next logic step (Score:4, Interesting)
Nah, there was definitely some classes in high school that had you just given me the book and assignments, required no input from the teacher. In fact, at least one of my history teachers and one of my math teachers was more harmful then helpful in the process of learning. Heck, one of my science teachers wasn't much better (but she was really hot, straight out of college so all was good).
Some of us learn better from reading and writing, others do better listening and talking. I am much better off reading information then listening to some speak the same information. Writing it down reinforces the learning where as just repeating what was said to me might not stick as well. I can see this in myself specifically in that I tend to repeat myself if I feel the person I'm talking to is not understanding or doesn't seem to be. This might be called mansplaining, but then again men do this to each other in abundance. I like to think of it more as confirming as much to myself as to others that I do in fact know what I am talk about. If I am incorrect, point it out for crying out loud.
Re: (Score:2)
Re: (Score:2)
AI may be able to provide information
Much of it completely false...
It cannot provide the personalized instruction, emotional support, and critical thinking skills that are necessary for a well-rounded education.
I couldn't agree more.
Re: (Score:2)
You must be one of those right-wing nuts that think teachers are "groomers" or forcing kids to change genders or whatever. I can't keep up with your insane conspiracies. Ugh...
ChatGPT will tell you lies. It will generate complete nonsense. This is not something that can be fixed. This is an unavoidable consequence of the technology.
You can not replace teachers with ChatGPT. That's absurd.
Re: Next logic step (Score:2)
Right place, right approach (Score:5, Interesting)
This class is all about new tech, innovation etc so his approach sounds great, especially since students are told to double check from other sources and are fully respond for AI errors.
Not necessarily appropriate for other fields which will vary dramatically across disciplines.
This reminds me of a scifi short story I read as a kid where everyone was "injected" with the knowledge to do some particular job in society but the protagonist "failed" the test and got dragged off to an isolated school the rest of the population was told was for the useless retarded kids. But it was actually where the next generation of scientists and engineers went who developed new technologies while everyone else only knew what was injected.
Someone still has to do original work even if we decide it's ok 99% of the population let's an AI think for them.
Provided has a cell phone line (Score:2, Informative)
Not necessarily appropriate for other fields which will vary dramatically across disciplines.
Reliance on ChatGPT is also not appropriate for less affluent students who don't come in with their own cell phone subscription. ChatGPT and other models offered by OpenAI require each user to verify ability to receive SMS [openai.com] at a unique phone number [openai.com] and don't accept landlines or VoIP [openai.com]. Even compared to a Texas Instruments graphing calculator sold at College Board cartel pricing, this might be a bit much because unlike a calculator purchase, a cell phone subscription incurs a recurring fee.
Re: (Score:3)
Not necessarily appropriate for other fields which will vary dramatically across disciplines.
Reliance on ChatGPT is also not appropriate for less affluent students who don't come in with their own cell phone subscription. ChatGPT and other models offered by OpenAI require each user to verify ability to receive SMS [openai.com] at a unique phone number [openai.com] and don't accept landlines or VoIP [openai.com]. Even compared to a Texas Instruments graphing calculator sold at College Board cartel pricing, this might be a bit much because unlike a calculator purchase, a cell phone subscription incurs a recurring fee.
I get where you are coming from, but you can also get a cheap cell phone and refill as you go service really cheap. You can get a basic flip phone for as little as $5 - 40 and unlimited talk and text for $15/month; if all you need is to get SMS and have a phone.
Re: Provided has a cell phone line (Score:2)
Re: (Score:3)
I do not have 15 dollars a month to spend, you are segregating me with these comments.
To help you desegregate, I would suggest getting a friend with an unlocked phone to try out one of the free trials TMOB et.al. offer since it comes with a new number to text and call. Unless, of course, you have no friends.
Re: Provided has a cell phone line (Score:2)
Then buy a prepaid sim card. That's a one-time $10 or $15 event, and essentially recharging it once a year with the smallest amount possible ($5-10). Then get the cheapest phone you can. Should be around $1.
If that doesn't work, as a last resort, try a free SIP account that receive SMS, e.g. Sipgate.
Re: (Score:2)
free SIP account that receive SMS, e.g. Sipgate.
From "Why am I not receiving my phone verification code?" [openai.com], with my emphasis:
Re: Provided has a cell phone line (Score:2)
Sipgate will assign you a phone number with your local area prefix, they're indistinguishable from a "real" landline. FWIW, most "real" landlines are provided to you through a SIP account, it's just one that you don't have the credentials to and which your ISP has preconfigured in the network device they're shoving down your throat.
Re: (Score:2)
Stop, poor families have so-called obama phones (though Bush started the program) and schools and libraries have computers. Is it unreasonable to expect students to go to the library to work on their assignments?
Why would a school populated with students that lack internet access at home hand out assignments that require internet access? Wouldn't the teacher give class time to work on the assignment in school?
Re: (Score:2)
Stop, poor families have so-called obama phones (though Bush started the program) and schools and libraries have computers. Is it unreasonable to expect students to go to the library to work on their assignments?
Why would a school populated with students that lack internet access at home hand out assignments that require internet access? Wouldn't the teacher give class time to work on the assignment in school?
Considering Wharton is an elite college at the undergrad and grad level; I suspect students have internet access through the school and are working on assignments at the school.
Re: (Score:2)
Is it unreasonable to expect students to go to the library to work on their assignments?
Yes. Not every kid has the luxury of staying after school to work on their assignments in the library. They might have to work, or be home to care for younger siblings so that someone else can work, for example.
Wouldn't the teacher give class time to work on the assignment in school?
Maybe, but probably not enough to satisfactorily complete the assignment. This unfairly disadvantages some kids who don't have the luxury of working on their assignments in the library after school or at home, due to other obligations or a lack of resources.
Re: (Score:2)
Why would a school populated with students that lack internet access at home
It's not that they lack Internet access as much as that they lack their own Short Message Service number.
Re: (Score:3)
Reliance on ChatGPT is also not appropriate for less affluent students who don't come in with their own cell phone subscription.
I think it's a pretty safe bet that any student going to Wharton already has a mobile phone.
Re: (Score:2)
That's not a fair assumption. Not everyone at an elite school is going to be an elite. I have no doubt that there is at least one scholarship student trying not to starve to death.
Re: (Score:2)
That's not a fair assumption. Not everyone at an elite school is going to be an elite.
You are so right [phillymag.com].
Re: (Score:2)
That's not a fair assumption. Not everyone at an elite school is going to be an elite. I have no doubt that there is at least one scholarship student trying not to starve to death.
A Wharton MBA is not a program for kids coming straight out of high school. You must already have a bachelors degree and have worked after college in a managerial capacity. Now I suppose somewhere in the multiverse that could include the rare poor person who doesn't already have a mobile phone, but it's fairly unlikely, hence why I said "a pretty safe bet."
Re: (Score:2)
Reliance on ChatGPT is also not appropriate for less affluent students who don't come in with their own cell phone subscription.
Oh, no! Life is unfair! Stop the world!
Re:Provided has a cell phone line (Score:5, Insightful)
Oh FFS, stop. Which schools don't have internet? Every 4 years we propose billions more in funding to "bridge the digital divide" - it must be under control by now.
And no, truly poor Americans qualify for free cellphones w/ no recurring fee, public schools have computers and internet, as do public libraries. The story of children being driven to McDonald's to use the free WiFi is BS. It's something some people did during Covid lockdowns because they were locked out of classrooms with computers and internet access, then the federal government threw another $100BN at the problem because it was election season.
The government is handing out cellphones, tablets, subsidized internet access, funding computers and internet access in schools, etc. it's time to admit the digital divide is shrinking, and we can't base every decision on the assumption that children lack access to the internet.
As a reminder, this article is about a graduate-level course at Wharton, not Miss Jones 3rd grade writing class. By the time this filters down to be part of the curriculum in K-12 public school education we'll have thrown countless billions more dollars at the "digital divide".
Land line to dorm is inadequate (Score:2)
Which schools don't have internet?
My complaint is not about Internet. I had serviceable Internet in my dorm when I attended college from 1999 to 2003. It's about cell phones. The only phone I had then was a land line, and a land line cannot verify an OpenAI account.
Re: Land line to dorm is inadequate (Score:2)
It was 20 years ago, maybe it's time to let it go?
The issue is now, not your childhood.
If a school lacks internet access in the classroom and in the library, it not for a lack of federal funds to make it happen - it's either because the school can't figure out the grant application process or they simply lack interest in getting internet access for the students.
Do you imagine any school front office lacks internet access to file federal attendance, free lunch program reports? That student grades are typed u
Again, I'm not talking about Internet. (Score:2)
Again, I'm not talking about Internet. I'm talking about SMS. Last I checked, a school provided Internet to dorms and didn't provide SMS.
Re: Provided has a cell phone line (Score:2)
Competitors are free to offer alternative graphing calculators to schools, they just need to invest in an ecosystem like the one that has evolved around the TI graphing calculator.
School districts buy the calculators by the pallet and loan them out to students - in many states (all?) it is illegal for a school district to require a student to buy something to complete a school assignment - a teacher can't 'require' a student to buy a TI graphing calculator to complete an assignment.
Nope. (Score:2)
You have to register an account and LINK A PHONE NUMBER to even create it. Nice job handing your kid's data to a "non profit".
OpenAI is an Orwellian name. If you want to do something get the open source shit like stable diffusion code or something else. This is absurd and dangerous.
Re: (Score:1)
If you want to do something get the open source shit
it is open source and available on github
Re: (Score:2)
Are you referring to this? [github.com] Which is literally just a front end for the website with the same requirements?
Re: (Score:1)
Re: (Score:2)
I'm not a progammer. From the 2nd link I don't see any code though, only "models" and from their own text it says:
s (51 sloc) 8.99 KB
GPT-3 Model Card
Last updated: September 2020
Inspired by Model Cards for Model Reporting (Mitchell et al.), weâ(TM)re providing some accompanying information about the 175 billion parameter GPT-3 model.
Model Details
GPT-3 is a Generative Pretrained Transformer or âoeGPTâ-style autoregressive language model with 175 billion parameters. Researchers at OpenAI developed the model to help us understand how increasing the parameter count of language models can improve task-agnostic, few-shot performance. Once built, we found GPT-3 to be generally useful and thus created an API to safely offer its capabilities to the world, so others could explore them for commercial and scientific purposes.
Model date
September 2020
Model type
Language model
Model version
175 billion parameter model
Paper & samples
Language Models are Few-Shot Learners
Release repository containing unconditional, unfiltered samples (CONTENT WARNING: GPT-3 was trained on arbitrary data from the web, so samples may contain offensive content and language.)
Model Use
The intended direct users of GPT-3 are developers who access its capabilities via the OpenAI API. Through the OpenAI API, the model can be used by those who may not have AI development experience to build and explore language modeling systems across a wide range of functions. We also anticipate that the model will continue to be used by researchers to better understand the behaviors, capabilities, biases, and constraints of large-scale language models.
Given GPT-3â(TM)s limitations (described below), and the breadth and open-ended nature of GPT-3â(TM)s capabilities, we currently only support controlled access to and use of the model via the OpenAI API. Access and use are subject to OpenAIâ(TM)s access approval process, API Usage Guidelines, and API Terms of Use, which are designed to prohibit the use of the API in a way that causes societal harm.
We review all use cases prior to onboarding to the API, review them again before customers move into production, and have systems in place to revoke access if necessary after moving to production. Additionally, we provide guidance to users on some of the potential safety risks they should attend to and related mitigations.
This sounds like you still have to register, provide a phone #, etc to actually any of this.
Re: (Score:2, Interesting)
You can't expect poor people to handle money appropriately!!!
Joke I grew up very poor, even on welfare and I honestly think it helped me really appreciate the value of earning a buck and how to live within ones means. Welfare wasn't THAT bad but the food got real slim at the end of the month. The welfare apartment was actually much nicer then the next several places we lived once we were paying our own way. Go figure.
Re: (Score:1)
Re: (Score:1)
This reminds me of a scifi short story I read as a kid
It reminds me of a story I read when I was about 10 years old to - "Danny Dunn and the homework machine" - published in 1958. In that book the protagonist and his friends got access to an advanced computer that could do their homework for them - including writing essays as well as maths. The teacher wised up to it and instead of banning them from using it, just started setting more challenging assignments with the expectation that they would use the the machine to complete, but would still need to know how
Re: (Score:2)
That was Asimov, and ChatGPT tells me the title was "The Fun They Had".
Re: (Score:2)
Hah, yes, sort of, but a little more direct in the story since the people literally got an injection. It was written in the 50s or 60s so the tech very hand wavy but same idea.
Uh, no (Score:5, Insightful)
I haven't even used the thing yet ... because it requires me to register and give it personal information. I should have to do that just to try it out? Seriously?
And we're going to make students in school do that? Seriously?
Schools are just clueless. Of course, many of them have been making kids have Google accounts for years, so I'm not sure why I am surprised.
Re: (Score:2)
I haven't even used the thing yet ... because it requires me to register and give it personal information. I should have to do that just to try it out? Seriously?
I guess you also shun mobile phones for the same reason. Or any utility contract?
Re: (Score:1)
You are comparing utilities (basic essentials to out modern life) to some "free" software online? Nice apples to oranges comparison there.
Re:Uh, no (Score:4, Insightful)
Then why are you complaining? You have the choice of not using it.
Either it's essential, in which case you don't have a choice, or it's optional, in which case you can ignore it.
Homework vs exams (Score:5, Insightful)
Re: (Score:2)
We use to have to write stuff down on paper. So even if a relative was helping us, we still had to be the one to write the stuff on the paper. The act of writing something down does help commit it to memory because you have to think about what you are writing. Sure, that alone won't make you proficient but it helps.
get along with? (Score:3)
I don't think Ethan Mollick's message is for the machines. The author starts off the first sentence like it's the machines we're supposed to be getting along with. Then immediately switches to the real subject of what Ethan Mollick is talking about there after - ChatGPT is a useful, albeit unreliable, tool. Document how it was used when authoring anything with its input, so the steps and results can be replicated.
Re: (Score:1)
These people need to get a clue too. ChatGPT is not AI. It's simply an advanced search engine.
It's not alive. It's not self aware. It's not creative. It can't think. It just organizes and spits out existing knowledge related to the terms you type in.
Expert Systems were initially developed in the 70's though their hierarchical nature limited their ultimate usability. This is just that except on a larger scale and more free-form.
Re: (Score:2)
I'm sure I've seen it posted a time or ten on here that many jobs could be replaced with a clever bash script. TRUE AI doesn't exist yet, but I guess it's how you want to define AI.
I think you are correct it's not AI, but a sophisticated bash script IS artificial intelligence to the typical human.
Re: (Score:2)
ChatGPT is entirely dumb. It never has a clue if it is right or not.
Since when were we in an A.I. world? (Score:2, Insightful)
Maybe adopting the tool instead of fighting it... (Score:4, Insightful)
... is the way to go. I still remember (vividly) when using a calculator was considered "cheating" even during my university years (late '80s to early '90s).
Having said that:
ChatGPT could be a serious disruptor to EVALUATION, if your evaluation relies heavily on papers, reports, thesis and such.
As some of my favourite MBA teachers ( @PabloMartindeHolan & @Bill Carney ) have said, in short succession, the potential for cheating is outstanding by using ChatGPT to write term papers, essays, or even answers to remote exam questions.
When I was an Engineering teacher, 60% of the grade was presential exams, and 60% of the grade* was a term-long practical project, so, no opportunity to use chatgpt to cheat. But I fear for the humanities or Magament type (empresariales) careers if professors do not adapt (like Ethan here did)...
But, at the same time, ChatGPT could AID the study process. For instance, in my field, you could use ChatGPT to study more efficently by asking it to condense the themes instead of reading the books, or even feed it past exams of mine (readily available, with my blessings, and answers) to see the answers. For instance, you feed it the relevant chapters of the Tanembaum, the Stallings, my slides, and my past exams and ask it to develop a study guide and some practice exams for you.
For the time being, presential exams and interactive oral debriefings are the way I see if you want to employ the current model of evaluation, or changes and adaptations, like Mr. Ethan here. In the future, who knows.
Tough times ahead.
Re: (Score:3)
When I was an Engineering teacher, 60% of the grade was presential exams, and 60% of the grade* was a term-long practical project, so, no opportunity to use chatgpt to cheat.
Please, tell me more about the Engineering school where the teachers aren't required to have a functional grasp on the concept of percentages... LOL
Re: (Score:2)
Nah, his engineering class just made them do 120% of the work.
Re: (Score:2)
Nah, his engineering class just made them do 120% of the work.
60%+60%= 120% on purpose, That way, students get a shot o achievenig 100%, even if they make little mistakes here and there. The mistake was me forgeting to put this note in the * at the botom of the comment
Re: (Score:2)
When I was an Engineering teacher, 60% of the grade was presential exams, and 60% of the grade* was a term-long practical project, so, no opportunity to use chatgpt to cheat.
Please, tell me more about the Engineering school where the teachers aren't required to have a functional grasp on the concept of percentages... LOL
60%+60%= 120% on purpose, That way, students get a shot o achievenig 100%, even if they make little mistakes here and there. The mistake was me forgeting to put this note in the * at the bootom of the comment
"I think everybody is cheating" (Score:4, Funny)
Perhaps at Wharton, everybody IS cheating ... which, given the prestige that school has in the business world, might explain a lot.
Finally some news that's positive! (Score:3)
This is a great idea on the professors part. Make them use that tool and then they can refine the output and improve upon that. Saves time but also helps them learn and still requires some thinking when they are cleaning things up and making it their own.
While AI might not kill all the jobs, it will allow one person to do much more in a given time frame where in the past that same amount of work in that same amount of time would of required more people involved.
So instead of needing 3 people you can get away with 1. From a top down perspective, that's freeing up human resources for other productive activities. It's still scary but that's because of our dog eat dog world. If we were not so hyper competitive this could be truly uplifting. It still might be.
The normalizing of idiocy... (Score:1)
Re: (Score:2)
The new Wikipedia (Score:3)
I would be interested to see someone compare ChatGPT outputs against well formatted Google searches, to see if you can tell where ChatGPT is pulling its data from. Something like ‘summarize Mark Twain’s Huckleberry Finn’, should allow one to see if the ChatGPT output simply plagiarizes portions of the google results, or if it has more going on.
Either way, I don’t see the value in it. If the output from it is really ‘original’ we are simply encouraging people to further offload their creative thinking skills, and if the output is just parroting other peoples work, the same problem applies, offloading critical thinking skills and creativity, and just using other peoples work.
I think there is a lot to be said for the idea that school is not so much about learning particular things, but training a person to be able to understand problems, formulate and ask questions, think of solutions, and solve problems; and it seems like this tool helps circumvent that.
Every advance brings this issue (Score:2)
Before literacy, there were village elders who *remembered* things. We tend not to have that now, because we can write things down; but writing didn't kill memorization. It survived (at least in my childhood) in the form of reciting poetry, memorizing state capitals, etc.
Same deal with photography. It didn't kill painting. In some ways it made it more interesting--encouraging artists to move beyond realistic portraits and landscapes in to abstraction and such... it may not be your taste, but that's OK b
Fair enough (Score:2)
However, it did leave me wondering why we needed the students at all.
Idiotic to compare ChatGPT to a calculator (Score:2)
Being able to write clear, understandable prose is a deep skill. Not everyone has it, and people who do have it take years to develop the skill (in an ideal world, the development happens when you are a teenager).
Also, reading someone's prose tells you a great deal about their thinking process. You can tell if they are precise and logical thinkers, you can tell whether they are biased towards a certain outcome, and you can tell whether they are trying to hide or obscure something. (Orwell talks about thi
Opportunity (Score:2)
Better keep learning (Score:1)
Well (Score:2)
School of plagiarism and highly derivative work (Score:2)
We can still test accuracy and comprehension (Score:2)
This approach makes a lot of sense to me. First off, we already know that AI systems don't always give you precisely the right answer. If a student uses an AI that provides them with the wrong answer -- even if only in part -- and that student doesn't catch and correct the mistake, then it seems to me that they've already failed on that assignment.
What's more, while we may appear to be reducing our testing of the writing skills of these students, we can still test their comprehension of the material by way
Calculators (Score:1)