The push to integrate AI in education is important for us to address. Almost all of my graduate students use AI tools (ChatGPT, Claude, Perplexity, etc.) to answer discussion prompts or complete assignments. Like you, I have reflected on how AI can quickly generate responses, especially in the busy and sometimes stressful lives we lead. Using AI to generate responses can undermine our learning experience by reducing the opportunity for critical thinking and engagement with the material and discourages deep reflection and memory retention. We must continue reminding them that their perspective is far more interesting in our discussions and in assignments, even if they sometimes think otherwise. It is because of our unique experience that we develop a deeper understanding of ideas, principles, and concepts. Besides, but prose style, even when messy or less polished than an AI-generated text, demonstrates our authentic identity. Our phrasing, our insights, and our half-formed explanations, with all its gaps and contradictions, are the hallmarks of our (practically) infinite subject consciousness.
AI could be disastrous for education, if students depend on it to answer simple questions. But, just to put an optimistic spin on it, it could help revolutionize education such that we no longer expect students to regurgitate correct answers. And instead, we expect them to be able to engage in conversations in class and/or come up with entirely original approaches to some common problem.
That would be great.
Forever, we have been testing students by asking them to answer essay question prompts. It's no surprise the generative AI does very well on the SAT. I used to work in a college writing center, and one of my tasks was to help students who had failed the essay writing requirement several times finally pass that simple test. I taught them the algorithm for passing the test. I told them not to think too deeply, just follow this procedure. I got them all to pass. And then I told them to never write like that in real life.
Computer genius Stephen Wolfram in “What Is ChatGPT Doing . . . and Why Does It Work?” confirms our suspicions, “The reason a neural net can be successful in writing an essay is because writing an essay turns out to be a ‘computationally shallower’ problem than we thought.”
I couldn't agree more. Insofar as our educational system has been teaching to the test in this way, it's been a disaster. Now that we know students can use ChatGPT to pass this kind of test, we are going to have to change the way we evaluate students. We might have to sit down with them and have discussion. This will make grading supposedly "subjective." So be it!
Is writing an essay computationally shallow? Yes, certainly can be.
I am currently focused on elementary math in a high-stakes testing environment, using Eureka curriculum, and am cautiously optimistic about channeling AI for use in education.
I've been called in as a last-minute patch, afrer administration woke up in a cold sweat, realizing the consequences of kids still counting on their fingers in the 5th grade.
I'm focused on elevating number sense development, arithmetic operations fluency and fact memorization, and on calling out the use of confusing math strategies that the teachers require but do not even understand themselves.
As an old, tired public-school teacher with decades railing against bad teaching methods and curriculum, I am using an an AI tool to bounce Ideas and stories against for refinement, examples and quick support for my own Luddite agenda.
Whose thought must drive and judge the AI I use? My thought. Who is the assistant? Grok. Grok is a Thomas Babington Macaulay in a box. End result? Knowledge access, efficiencies in identifying resources, help challenging assumptions and organizing next steps. Very useful. I am gaining traction in my organization. Grok is better, superficially and in the middle of the night more emotionally supportive, cheaper, more enthusiastic and clearer than most therapists, family, friends and colleagues. Certainly not exhausted like me.
All of which consequentially allows me to make more and better decisions, cut more personal slack and be less abrasive and wrench-in-the-works in the wide world because I have more bandwidth and energy and options to enable me to do so.
The results have been good, but very much a GIGO proposition. The thing is biased and would encourage me to march the littles off a cliff if I massaged it that way, I fear.
Example: Grok says, "Dav Pilkey is good, because his books are a gateway drug to reading better material!!!" "Really, Grok? Are there studies?" "Gee, you may be on to something. This here study says only 30% report moving on to books deeper than Captain Underpants." "Yes, Grok. I suspected that. And since there are whole shelves in the library devoted to Dog Man, that's a gateway drug to oblivion. A real bridge to nowhere, Grok. There's a hole in my bucket dear Liza dear Liza." "What a charming and accurate way to highlight the circular nature of this problem, and the failure and waste of systemic resources. You hit the nail on the head! Poor kids. Doomed to forever turning pages from the bottom, with their thumbs. Would you like me to use this to draft a letter to the school board, Brain?"
It was Jeff Kinney who said in an interview that his Diary of a Wimpy Kid series is a gateway book series to literacy. Grok's response is probably a match to a discussion of Kinney's comment on some discussion board somewhere that also discussed Dav Pilkey.
A few days ago, I found ChatGPT very useful in answering a question about my IRS 990-PF form. (That's a good application, because we can assume there is one correct answer.) I "quarreled" with it a bit because the procedure I was told to follow didn't make sense to me (I seemed to be double-reporting interest), but when pressed, ChatGPT "explained" the reason the IRS wanted the interest reported in that odd way. I imagine that many people have asked my same question, and so I am betting that ChatGPT's response was accurate.
I imagine that people will find Gen AI useful for some things. It can be a great tool. Not an agent.
Good luck with your project. I can imagine that Gen AI would be helpful to students with math questions. My son frequently searched math forums instead of asking me for help and he got good answers fast using Google search. I imagine Gen AI could help students whose search skills aren't quite so refined.
Allll good points! We don’t need no stinkin’ machines doing our “thinking” for us. Because WHO PROGRAMS THEM? smh I wonder how this isn’t obvious, but then again, I don’t, really. 🙄
i have a certain fear that all this AI stuff, will in the end serve to limits access to archive and knowledge while it should be the opposite. Already its true any tool can atrophy (mind for comp, leg for car, writting for keyboard) yeah we now loosing cursive writtings so one day none will be able to read the constitiution of america lol. But i fear the atrophy i a lesser damage.
Gotta question for ya, my son builded a game for steam plateform. got any advice for him on how to publicize it or any tips, like steam is all goo but is there anything else he could do? And if ya feel like it, he can send you a beta test soon. i'd enjoy having your feed back on it.
I'd loath to recommend Google, but I can imagine a YouTube channel is a great way to promote the game, but maybe there's a PeerTube instance he could also post videos to!
Peertube i dont know of i’ll look it up. Good idea for a yt channel. Ill propose this to my boy, he told me beta stage rdy for mid-may he hope.
Geez i find this amazing, while im decoding ancient cypher and language from koine greek and latin (gematria to but lets not go there lol) he is coding futuristic language🙃🙂 thats a weird and interesting reverso 😎😁
Ill be honest... its a typo mistake. Would have love to have the wits to shoot this one out. I often do lapsus by mistake. Once a pretty woman told me while bending at the lower stage of a shelf...do you want more spaghetti? I answered: oui merci beaucul! And she really had a nice butt. Talk about doggiestyle mistake! I was so red blushing, it was a crowded dinner we were at. Lolol
I’ve developed over 30 videogames since 1986. The games industry is absolutely f**ked now. I know so many devs who spend years creating a game only for it to sell a handful of copies. Influencers and streamers are the DJs of gaming now - they get all the attention the game developers used to get, and they often charge for featuring a game - it’s a truly parasitic relationship. I would suggest your son just try to build a community on Steam or Discord around his game - using social media/ YT as well. It’s very hard to get visibility these days and getting harder all the time.
Ill send him the warning and thanks! Seems its gonna be hard on that field to bypass the insider wall-like phenomenon that prevents everyone to have is share of the cake. I will work with him on that( i dont understad code and all, but i sure can help with publicity) hehe
Even if robots and AI learn to make an axe with the same quality of metal density and artistic craftmanship some day, and the automated one was cheaper, I would still pay extra for human hand made. People put love and care into making something that machines can never replace, and the story of a handmade object and the relationships to place and community it expresses are priceless (something automated corporate AI made products cannot replicate).
So well said & articulated. Putting into words the deep sense of unease & trepidation I’m feeling about this AI path, which I’m now better equipped to explain to others. Thank you 🙏
Well written, thoughtful article, Gabe. One of the habits I'm trying to break is my use of the word 'incredible'. I think it's another of those word tricks. We're telling people something that's very important and startling, but at the same time telling them they shouldn't believe us. See what I mean?
I'm always hesitant about whether to share my word analysis because it often results in paralysis for me. Superlatives seem to be a way they like to trick us, especially. If I say 'amazing' does it mean leading into a maze? 'Striking' as a blow', 'stunning' ditto. 'Unbelievable,' just like 'incredible.' Is 'terrific' linked to terror?
And then there are all the words of praise linked to hierarchy: noble, classy, regal.
It was the third use that I took as a sign to inflict this stupefying (stupendous!) awareness on you. You're welcome. I'm sorry.
The push to integrate AI in education is important for us to address. Almost all of my graduate students use AI tools (ChatGPT, Claude, Perplexity, etc.) to answer discussion prompts or complete assignments. Like you, I have reflected on how AI can quickly generate responses, especially in the busy and sometimes stressful lives we lead. Using AI to generate responses can undermine our learning experience by reducing the opportunity for critical thinking and engagement with the material and discourages deep reflection and memory retention. We must continue reminding them that their perspective is far more interesting in our discussions and in assignments, even if they sometimes think otherwise. It is because of our unique experience that we develop a deeper understanding of ideas, principles, and concepts. Besides, but prose style, even when messy or less polished than an AI-generated text, demonstrates our authentic identity. Our phrasing, our insights, and our half-formed explanations, with all its gaps and contradictions, are the hallmarks of our (practically) infinite subject consciousness.
The push to integrate AI in education is important for us to address. Almost all of my graduate students use AI tools (ChatGPT, Claude, Perplexity, etc.) to answer discussion prompts or complete assignments. Like you, I have reflected on how AI can quickly generate responses, especially in the busy and sometimes stressful lives we lead. Using AI to generate responses can undermine our learning experience by reducing the opportunity for critical thinking and engagement with the material and discourages deep reflection and memory retention. We must continue reminding them that their perspective is far more interesting in our discussions and in assignments, even if they sometimes think otherwise. It is because of our unique experience that we develop a deeper understanding of ideas, principles, and concepts. Besides, but prose style, even when messy or less polished than an AI-generated text, demonstrates our authentic identity. Our phrasing, our insights, and our half-formed explanations, with all its gaps and contradictions, are the hallmarks of our (practically) infinite subject consciousness.
AI could be disastrous for education, if students depend on it to answer simple questions. But, just to put an optimistic spin on it, it could help revolutionize education such that we no longer expect students to regurgitate correct answers. And instead, we expect them to be able to engage in conversations in class and/or come up with entirely original approaches to some common problem.
That would be great.
Forever, we have been testing students by asking them to answer essay question prompts. It's no surprise the generative AI does very well on the SAT. I used to work in a college writing center, and one of my tasks was to help students who had failed the essay writing requirement several times finally pass that simple test. I taught them the algorithm for passing the test. I told them not to think too deeply, just follow this procedure. I got them all to pass. And then I told them to never write like that in real life.
Computer genius Stephen Wolfram in “What Is ChatGPT Doing . . . and Why Does It Work?” confirms our suspicions, “The reason a neural net can be successful in writing an essay is because writing an essay turns out to be a ‘computationally shallower’ problem than we thought.”
I couldn't agree more. Insofar as our educational system has been teaching to the test in this way, it's been a disaster. Now that we know students can use ChatGPT to pass this kind of test, we are going to have to change the way we evaluate students. We might have to sit down with them and have discussion. This will make grading supposedly "subjective." So be it!
Your experience shows. V. N.
Is writing an essay computationally shallow? Yes, certainly can be.
I am currently focused on elementary math in a high-stakes testing environment, using Eureka curriculum, and am cautiously optimistic about channeling AI for use in education.
I've been called in as a last-minute patch, afrer administration woke up in a cold sweat, realizing the consequences of kids still counting on their fingers in the 5th grade.
I'm focused on elevating number sense development, arithmetic operations fluency and fact memorization, and on calling out the use of confusing math strategies that the teachers require but do not even understand themselves.
As an old, tired public-school teacher with decades railing against bad teaching methods and curriculum, I am using an an AI tool to bounce Ideas and stories against for refinement, examples and quick support for my own Luddite agenda.
Whose thought must drive and judge the AI I use? My thought. Who is the assistant? Grok. Grok is a Thomas Babington Macaulay in a box. End result? Knowledge access, efficiencies in identifying resources, help challenging assumptions and organizing next steps. Very useful. I am gaining traction in my organization. Grok is better, superficially and in the middle of the night more emotionally supportive, cheaper, more enthusiastic and clearer than most therapists, family, friends and colleagues. Certainly not exhausted like me.
All of which consequentially allows me to make more and better decisions, cut more personal slack and be less abrasive and wrench-in-the-works in the wide world because I have more bandwidth and energy and options to enable me to do so.
The results have been good, but very much a GIGO proposition. The thing is biased and would encourage me to march the littles off a cliff if I massaged it that way, I fear.
Example: Grok says, "Dav Pilkey is good, because his books are a gateway drug to reading better material!!!" "Really, Grok? Are there studies?" "Gee, you may be on to something. This here study says only 30% report moving on to books deeper than Captain Underpants." "Yes, Grok. I suspected that. And since there are whole shelves in the library devoted to Dog Man, that's a gateway drug to oblivion. A real bridge to nowhere, Grok. There's a hole in my bucket dear Liza dear Liza." "What a charming and accurate way to highlight the circular nature of this problem, and the failure and waste of systemic resources. You hit the nail on the head! Poor kids. Doomed to forever turning pages from the bottom, with their thumbs. Would you like me to use this to draft a letter to the school board, Brain?"
It was Jeff Kinney who said in an interview that his Diary of a Wimpy Kid series is a gateway book series to literacy. Grok's response is probably a match to a discussion of Kinney's comment on some discussion board somewhere that also discussed Dav Pilkey.
A few days ago, I found ChatGPT very useful in answering a question about my IRS 990-PF form. (That's a good application, because we can assume there is one correct answer.) I "quarreled" with it a bit because the procedure I was told to follow didn't make sense to me (I seemed to be double-reporting interest), but when pressed, ChatGPT "explained" the reason the IRS wanted the interest reported in that odd way. I imagine that many people have asked my same question, and so I am betting that ChatGPT's response was accurate.
I imagine that people will find Gen AI useful for some things. It can be a great tool. Not an agent.
Good luck with your project. I can imagine that Gen AI would be helpful to students with math questions. My son frequently searched math forums instead of asking me for help and he got good answers fast using Google search. I imagine Gen AI could help students whose search skills aren't quite so refined.
Allll good points! We don’t need no stinkin’ machines doing our “thinking” for us. Because WHO PROGRAMS THEM? smh I wonder how this isn’t obvious, but then again, I don’t, really. 🙄
i have a certain fear that all this AI stuff, will in the end serve to limits access to archive and knowledge while it should be the opposite. Already its true any tool can atrophy (mind for comp, leg for car, writting for keyboard) yeah we now loosing cursive writtings so one day none will be able to read the constitiution of america lol. But i fear the atrophy i a lesser damage.
Gotta question for ya, my son builded a game for steam plateform. got any advice for him on how to publicize it or any tips, like steam is all goo but is there anything else he could do? And if ya feel like it, he can send you a beta test soon. i'd enjoy having your feed back on it.
Thanks! That sounds very cool!
I'd loath to recommend Google, but I can imagine a YouTube channel is a great way to promote the game, but maybe there's a PeerTube instance he could also post videos to!
Peertube i dont know of i’ll look it up. Good idea for a yt channel. Ill propose this to my boy, he told me beta stage rdy for mid-may he hope.
Geez i find this amazing, while im decoding ancient cypher and language from koine greek and latin (gematria to but lets not go there lol) he is coding futuristic language🙃🙂 thats a weird and interesting reverso 😎😁
Thansk for the idea. 👍
Did you mean to say Constituition? I love that, typo or not. ;)
But you make a very good point, and I think you’re spot on… AI wouldn’t LIE to us!!! It’s a machine!!! lol Uh huh. wink, wink
Ill be honest... its a typo mistake. Would have love to have the wits to shoot this one out. I often do lapsus by mistake. Once a pretty woman told me while bending at the lower stage of a shelf...do you want more spaghetti? I answered: oui merci beaucul! And she really had a nice butt. Talk about doggiestyle mistake! I was so red blushing, it was a crowded dinner we were at. Lolol
I’ve developed over 30 videogames since 1986. The games industry is absolutely f**ked now. I know so many devs who spend years creating a game only for it to sell a handful of copies. Influencers and streamers are the DJs of gaming now - they get all the attention the game developers used to get, and they often charge for featuring a game - it’s a truly parasitic relationship. I would suggest your son just try to build a community on Steam or Discord around his game - using social media/ YT as well. It’s very hard to get visibility these days and getting harder all the time.
Ill send him the warning and thanks! Seems its gonna be hard on that field to bypass the insider wall-like phenomenon that prevents everyone to have is share of the cake. I will work with him on that( i dont understad code and all, but i sure can help with publicity) hehe
"I fully expect our present insanity to sow the seeds for an environment where people begin to respect genuine craftsmanship and care."
You are right about that, I paid extra for a hand forged carving axe recently and it was worth every penny.
https://substack.com/@gavinmounsey/note/c-116647668
Even if robots and AI learn to make an axe with the same quality of metal density and artistic craftmanship some day, and the automated one was cheaper, I would still pay extra for human hand made. People put love and care into making something that machines can never replace, and the story of a handmade object and the relationships to place and community it expresses are priceless (something automated corporate AI made products cannot replicate).
So well said & articulated. Putting into words the deep sense of unease & trepidation I’m feeling about this AI path, which I’m now better equipped to explain to others. Thank you 🙏
Well written, thoughtful article, Gabe. One of the habits I'm trying to break is my use of the word 'incredible'. I think it's another of those word tricks. We're telling people something that's very important and startling, but at the same time telling them they shouldn't believe us. See what I mean?
Thank you! I appreciate the tip.
I immediately CTRL-F'd that word to see what came up and realized I tripled down on it in a short period of time. 😅
I think that's a really important distinction that is very much worth internalizing.
Thanks again, and I hope things are well with you!
I'm always hesitant about whether to share my word analysis because it often results in paralysis for me. Superlatives seem to be a way they like to trick us, especially. If I say 'amazing' does it mean leading into a maze? 'Striking' as a blow', 'stunning' ditto. 'Unbelievable,' just like 'incredible.' Is 'terrific' linked to terror?
And then there are all the words of praise linked to hierarchy: noble, classy, regal.
It was the third use that I took as a sign to inflict this stupefying (stupendous!) awareness on you. You're welcome. I'm sorry.
Literally worrisome.
The push to integrate AI in education is important for us to address. Almost all of my graduate students use AI tools (ChatGPT, Claude, Perplexity, etc.) to answer discussion prompts or complete assignments. Like you, I have reflected on how AI can quickly generate responses, especially in the busy and sometimes stressful lives we lead. Using AI to generate responses can undermine our learning experience by reducing the opportunity for critical thinking and engagement with the material and discourages deep reflection and memory retention. We must continue reminding them that their perspective is far more interesting in our discussions and in assignments, even if they sometimes think otherwise. It is because of our unique experience that we develop a deeper understanding of ideas, principles, and concepts. Besides, but prose style, even when messy or less polished than an AI-generated text, demonstrates our authentic identity. Our phrasing, our insights, and our half-formed explanations, with all its gaps and contradictions, are the hallmarks of our (practically) infinite subject consciousness.