Teaching in the Age of AI

Dr. Josh Stout discusses how AI is changing student's work, but not in the way we expected.

Teaching in the Age of AI
AI interpretation of the mindbodyevolution baobab

AI is changing student's work, but not in the way we expected.

Teaching in the Age of AI
Dr. Josh Stout

Eric 0:02

It is Friday, February 23rd, and this is season two, Episode eight of Mind Body Evolution. Hi, Josh. 

Dr. Josh Stout 0:09

Hi, Eric.  So I've been thinking a lot about A.I. recently, and… 

Eric 0:16

You and everyone else. 

Dr. Josh Stout 0:18

Me and everyone else. And I've. I've been to a couple of professional development events at my university. I have. I spent some time looking at the products out there. I've been I've been thinking about it to the point where I we came up with the title of today's podcast using A.I. because it was better than our our titles. It worked better. 

Eric 0:43

Well, we were we were working with some clunky compositions. 

Dr. Josh Stout 0:45

Exactly. And so he just came out with a smoother, more professional sounding title than we would have come out with on our own. Yes. That is what it does. 

Eric 0:53

Yes, it does. 

Dr. Josh Stout 0:54

And it rubs the. All the all the all the bumps off of things and makes everything look nice and shiny, but also slightly soulless. And I think the title reflects that as well. It sounds exactly like a very, you know, typical Internet kind of title on a on an article or a podcast. 

Eric 1:10

Parenting and teaching in the age of AI. Yeah, it's very. 

Dr. Josh Stout 1:13

Very, very slick and slightly soulless. And that that's, you know, really the essence of AI. And that is, I think it's true dangers and it's, you know, true advantages and why we're all going to be using it. 

Eric 1:24

All that and the fact that it lies to you and presents it as truth. 

Dr. Josh Stout 1:27

Of all of these issues. 

Eric 1:29

You know. 

Dr. Josh Stout 1:29

Slickly and professionally. Yes. Yeah. And and, you know, can be directed to do that or will just do it on its own. Hard to say. Catch hell. Yeah. So it's really interesting being a professor as AI has been implemented and seeing sort of the iterations have happened. And just I want to start off by just talking about the writing and this is what everyone was first… 

Eric 1:52

You mean the writing that you are receiving from the students and how it has been affected by AI's being integrated. 

Dr. Josh Stout 2:01

Or technology in general up till now with A.I. being sort of the icing on the cake. What's changed recently? 

Eric 2:10

It's an interesting way of thinking of it because I don't see A.I. as icing. I see A.I. as a whole new technology that's changing everything the way print the printing press did it is. 

Dr. Josh Stout 2:19

But thinking about from the point of view of a student, a typewriter with a tremendous new advantage. Yes. A computer changed everything once. It could spell for you. I would not have gotten a Ph.D. if I didn't have a computer that could spell for me. 

Eric 2:35

I had a professor in college who told me to write a paper on a computer, and I fought it. And then it changed everything. It changed my life. 

Dr. Josh Stout 2:42

It changes everything. And it changes it because suddenly you don't have to worry about things like the spelling. It just does it for you. And it allows it frees you. 

Eric 2:51

Also frees you in other ways because hands writing is actually laborious, you can type much faster and suddenly thoughts become transformed into permanent words so quickly. 

Dr. Josh Stout 3:00

Absolutely all true. And so. 

Dr. Josh Stout 3:03

These were things that enabled me personally to succeed because I would not have been able to as a dyslexic person. I am really dyslexic. Really very, very dyslexic. I would not have been able to do anything. No, no one would have read my handwritten papers and the spelling would have made people think I was in fourth grade. So it was it was it was something that made it possible, you know, like like a wheelchair might make it possible for someone who can't walk to move around. It was it was really a new and incredibly useful tool. And I suspect that there are many people in my generation who were coming forward, you know, in the eighties and nineties, in the same kind of world where suddenly you realize that there is things you could do that you thought you could never do. And in science, to even larger extent, when I would go into the the the storage rooms in various science facilities, I like to look at the old stuff because I'm a curious person. I like to see cool things. And I would see these things that were called standard curves and they were pieces of plastic that had standardized curves for a graph you would be drawing. So if you wanted a bell curve, you could have a narrow one or a wide one and you would draw, you would hand draw it and submit that for publication. I would have never been able to publish anything, never a dissertation, nothing, because I cannot hand draw a graph. It doesn't matter if you give me a piece of plastic to trace, there's going to be blobs on it. There's going to be notches on it. It won't line up properly. There is absolutely no way I would ever, eve have been able to produce a graph. I started with with professors who'd actually trained in the old world. So when I first, you know, at first I'd gotten a master's in philosophy and then I switched to a PhD program and I had to get some undergraduate classes on the way. And so I was working with some of these older generations of professors who had done an awful lot of hand work for their papers. And one of the first things I had to do was take a print out in the United States and draw range maps for amphibians showing there where they are, where they where they where they used to be. And that kind of thing. 

Eric 5:19

You had to you had to freehand draw this. 

Dr. Josh Stout 5:21

I had to freehand draw a range map. 

Eric 5:23

So in order to be a scientist, you had to also be an artist? 

Dr. Josh Stout 5:26

Absolutely. Basically, in every aspect. So if you want if you needed an illustration, I had to be able to draw it. Pictures had actually been invented by the time the 1980s rolled around, so things were being photographed and such. But for a range map of organisms, those would be you'd have a printed out map and then you would draw on it where something was right. 

Eric 5:47

That range map didn't exist before you got the data and needed to create it. 

Dr. Josh Stout 5:50

And needed to create it. 

Eric 5:51

It. Exactly. Then it needed to be created. 

Dr. Josh Stout 5:53

And so you go back and forth between you would you would research sightings and you would put dots on the map and then you would draw a circle around those dots to show where something was. And I got a D on this project and there was there was no way I was going to get anything but a D on this project because my my hand was shaky as I did it. Things didn't line up properly. I would have needed like six copies of, of, of, of this this map to be able to do it. It was, it was a nightmare and it was never going to improve. But this would be tremendously easy on a computer today a map of the United States draw some circles in whatever program you're using. block them on the map. And it would be it would be taking no time at all where you were. 

Eric 6:39

You were able to do things you wouldn't have been able to do, as in it a handicapped person being prevented absolute in a society not built for them. 

Dr. Josh Stout 6:47


Eric 6:47

And so. So what are you seeing now? 

Dr. Josh Stout 6:50

Well, we're working up to that. And this is actually something I think about a lot is that if you build a world that helps the people who are having the most difficulty, you actually help everyone else as well. And so something that made it possible for me to exist, everyone else enjoyed being able to spell correctly as well and. 

Eric 7:11

And not have to hand draw... 

Dr. Josh Stout 7:12

And not have hand draw everything. So this, I think is is is is something where accessibility is something that we need to think about for everyone and should not be for particular individuals. That the better we make life for people who have difficulties, the easier it will be for all of us. You know, it's impossible for someone in a wheelchair to get upstairs.So  an elevator helps. An elevator will help everyone. 

Eric 7:37

I mean, this is this is the argument for for universal health care and universal basic income. I anyone think we're getting distracted. 

Dr. Josh Stout 7:44

But these are all two things that I directly connect with. The rise of technology and the way it's been interacting with education. And so I definitely am getting students who I wouldn't have seen before. Some of them have learning disabilities and they're they're using their computers to help them. Others have their English as a second language or they're first generation. None of none of their family have ever been to university. And so they don't speak the same standard English that a university would expect of them. And so this has been a bit of an issue for a while. A word after it was figuring out spelling started to work a little bit on grammar. And this this was helping students a little bit, but most of them were ignoring the grammar prompts. I was generally ignoring the grammar prompts because they seemed idiosyncratic. They didn't they weren't they weren't incredibly useful most of the time. And so I was getting papers that the spelling had largely been fixed. Sometimes I think the students don't even notice when they get the read underline on this on the word. You know, clearly there's typos. They just put two words together into one big word and they didn't notice and word did, but they didn't notice word noticing. So that's a separate issue that has nothing to do with, you know, computers or A.I., But this was something that had been helping students. But there was a limit to it. They their their English just couldn't get better with the existing technology. And I was often sending people to the writing center for things like simple, simple things like subject verb agreement. They, they, they, they would say, you know, the “people's is”, his instead of “the people are”. 

Eric 9:33

What is what is the writing center. 

Dr. Josh Stout 9:35

Or the writing center at the university where you just send people to be tutored and they would go through their paper and see what comments I had made and then understand what I meant because a lot of students would know what a subject verb agreement meant. 

Eric 9:48

Somewhere they can get help to interpret your comments. 

Dr. Josh Stout 9:50

Or interpret my comment because I'm interpreting the way a professor would interpret something saying, you know, your your plurals don't match your verbs and they need to and I can’t read your papers. 

Eric 10:01

So has the Writing Center become the AI? 

Dr. Josh Stout 10:05

Well, the writing center was definitely helping people. And then when AI came out, I was terrified. I had a lot of students who their English was not great and they had not come through a system writing in English, and some of them couldn't really write at all, even though they were in graduate classes with me. And they'd all gotten A's before because no one had ever really made them write much. And so I had decided I did that for my own sort of ability to see myself as a good teacher. I needed to make grad students write papers and this made them suffer. And some of them. 

Eric 10:49

That sounds like it would also make you suffer because you now have to teach them what you're teaching in addition to teaching them how to write. 

Dr. Josh Stout 10:54

Right.  And so I couldn't I couldn't teach them English, so I was sending them to the writing center so they could learn. I said, Look, you need to go every week and you need to work on this. And when they did this, I saw tremendous advancement. I saw them actually writing better papers somehow. 

Eric 11:08

I'm not surprised by that. 

Dr. Josh Stout 11:09

But some of them were trying to get around the system, not have to go to the writing center and not have to do the writing and so what they would do is they would cut and paste sections of a text. So if they had to write the methods section of a paper, it's just a list of techniques that were done in the actual paper and my students generally had to paraphrase it. But students who had a hard time with English were having a hard time paraphrasing First, it was difficult to read. The material is very technical, and then it was difficult to put it into your own words without just copying the technical information. And this was this was a struggle. So they would put it through something called a thesaurus or a text spinner, a thesaurus program. 

Eric 11:52

A text spinner? 

Dr. Josh Stout 11:53

Text spinner. And what it does is it it was a way of getting around safe. Assign. So safe assign is something the professors have. That's a record, basically, of all papers that have been sent to professors ever by professors. And so if a student just copies of a previous paper, it will say you are copying it and it tells you what percentage is copy. 

Eric 12:17

Where does this SafeAssign database? Get all the papers from, Do all the professors submit them? 

Dr. Josh Stout 12:23

When whenever I have students submit work, I click the little thing that says check with safe assign. And so in safe assign a second checking the paper. They're also recording the paper, essentially. 

Eric 12:35

So this is data that's that's collated from all the professors that are using it. 

Dr. Josh Stout 12:39

Yes. And also, I think it knows all the papers that have been, you know, put out in general. 

Eric 12:46

Forgive me. I know that most of the people listening to us are actually in this environment. I am not. Yeah. So I've been in the university environment. Yes. 

Dr. Josh Stout 12:55

Stepping back for a moment, this is for professors, these students, in many ways, they all should know what I'm talking. 

Eric 13:00

Well, what I’m asking is this is a like this is a ubiquitous thing now, and I just don't know about it because I haven't been there. Wow. Okay.

Dr. Josh Stout 13:05


Eric 13:06

So what? Sorry, everyone, for my ignorance.

Dr. Josh Stout 13:07

So what students had been doing is supposedly frats were doing this a lot. I'm not sure if it was all frats or not. I don't want to malign frats in general, but any any group of students with an institutional ability to save papers for next year right. People's younger brothers, this sort of thing would just give me the next paper again the next year. And it's very difficult to tell, especially when you have ten papers on the same subject. Getting one from the previous year would just be slipped in really easily. 

Eric 13:38

Forgive me, maybe it's just the Sarah Lawrence in me, but what's the point of not doing the work? 

Dr. Josh Stout 13:43

Exactly what most of the students really do the work. But there's always some who are seeking a way to get by. 

Eric 13:49

But you just so much work to not do the work. 

Dr. Josh Stout 13:51

I know exactly. And so I basically solved the problem with safe assign. And I was I was sending back papers to people. You know, this is showing up as 70% copy and you need to put this in your own words. And then what they would do is they would put it through a text spinner and a text spinner changes the orders and the words used. So that you have the same meaning, bu a different word and it a bit safer sign. But what it actually produces is gibberish, particularly for scientific papers, because it takes something like an immune system where you can't change those words. An immune system is an immune system and they would change it to protective framework. And so it would be reading these things. I could not understand what the students were doing and I would ask them and they'd say, Yes, I absolutely am writing this myself. I'm just using a little help from this. This text spinner to to, you know, get the language right. And I knew what they were actually doing was cutting and pasting chunks, then rewriting it. And if they rewrote it from the text better, it actually came out okay sometimes. But I could always tell because of the strange words that were coming out, very strange juxtapositions of language that didn't make any sense, didn't sound like a scientific technical paper, but they were just sort of getting by on this. 

Eric 15:04

So you could tell. You could tell on your own. 

Dr. Josh Stout 15:07

Yeah. And I got to the point where I was no longer going to accept this. I would say this in my syllabus. If I see any evidence of a text spinner, I'm not going to try and fix your gibberish. I'm just going to say try again. And I would have students drop out because if they had to try again, they couldn't really write English and so they had to leave. And I had students who tried again and got it and actually became good writers. It's amazing at it when you're in your early twenties. Simply just having someone write forced you to write ten papers over the course of the semester. The difference between the beginning and the end of the semester is amazing. I watched these students learn to write because no one's ever done this. 

Eric 15:47

When you make an effort, it pays off. 

Dr. Josh Stout 15:49

It really pays off. Yeah. I mean, most of the time when a student turns in a paper, they just get a grade on it. And so what I was doing was I was doing a system where they would turn in a paper. I would write comments and they would revise the paper for a new grade. And this really, really helped them learn how to do things, but it meant that they really had to do it. And that was a bar that some of them couldn't get over. 

Eric 16:14

Well, I mean, that sounds like a reasonable bar. 

Dr. Josh Stout 16:16

Yeah, absolutely. And now things have changed again. So we left behind the story Spinner world when we entered the world of AI. And I was afraid of what all the other professors were afraid of. I was afraid that the students were simply going to have I write their papers, turn it into me, and I'd see that it was too perfect, and I knew it was too good to be of students writing, but there would be nothing I could do about it. And I would. I would just have to give everyone an A and no one would be doing any work anymore. 

Eric 16:48

But that's not what happened. 

Dr. Josh Stout 16:49

It's not what happened. I had a pure writing class. It's all online. There's no reason that they couldn't do this. But because I got ahead of it, I realized what might happen. I wrote a long description in my syllabus of what I would like them to do. I said, You can turn in papers as long as you also give me the prompts. And I realized that looking at a scientific paper, you can't easily take a scientific paper and just cut and paste it into ChatGPT and then have it turn it out in the right way. 

Eric 17:21

Not a scientific paper. 

Dr. Josh Stout 17:22


Eric 17:24

But other kinds of papers, maybe you can. 

Dr. Josh Stout 17:27

But you're taking a journal article and then asking for a summary of it. You know, in a particular format. It was difficult for TVT. I was able to force chatty P to do it. If I broke it down into small segments and I could I could make each segment give me a good piece of text that made sense free of hallucination. If I kept it short. But I had to have a prompt to do that. And I told them, If you give me the prompts, I will approve any text you give me. But unlike if you write it yourself, there's a limit. You can't give me a 100 page paper and expect me to read it. I will read two pages of a I writing two single spaced pages of my writing. If you as a person give me four pages, a single spaced writing, I will read that. I will not read four single spaced pages of writing. You have to keep it in the limits because with air you can tell it what those limits are. And so I figured if I was teaching I engineering, I wanted to see prompts. They had to keep it to a very specific format and I wanted to see them do that. And so I thought that that would be really a useful way for students to learn how to use a I as a as a sort of writing tool. I hope they do this. I'm now on my third year of this two and a half, and no student has taken me up on this. Zero of them have actually used a I to write a paper when I have told them I will accept it. And you can you just have to tell me how you did it, because it's actually harder than what I was. 

Eric 18:59

I was going to say, it sounds like you're really integrating their work into what the AI is doing that way and you're making them prove it. I like this. 

Dr. Josh Stout 19:10

I was trying to. It just never happened. 

Eric 19:12

They just never… 

Dr. Josh Stout 19:13

They’ve never done it. 

Eric 19:13

So are they just submitting AI papers without telling you or are they doing the work themselves? 

Dr. Josh Stout 19:18

No, as far as I can tell, no. Whenever I see things improve too much or if there's if there's a if there's a strange, soulless feeling to a paper, I do run it through the AI detectors. 

Eric 19:31

What are the AI detectors? 

Dr. Josh Stout 19:34

They're programs that look for what's called bursting. Bursting is when a, the author of a paper chooses an unusual word or an unusual turn of phrase. And I don't do this. The eyes are very, very straight ahead. That's why they're soulless. They always write the same thing and sort of not the same way. They can actually write in a hundred different ways, but they will write it in that same soulless way every every time. And it's always going to be a very standard English. I'm not going to see any evidence that someone came from another country or someone's family came from another country. I'm not going to see unusual grammar formations. 

Eric 20:13

Or even interesting or unique or anything. 

Dr. Josh Stout 20:18

Exactly.  I will never throw in a comment right in the middle. I'll often tell them you really shouldn't in a scientific paper. How? But it lets me know I'm really reading a student. 

Eric 20:25

When you run it through a detector, how does it return that response? Does it? Is it just like, yes, this was written by and no, it wasn't. I mean, is it that clear? 

Dr. Josh Stout 20:34

No, it's actually really funny. It gives me a percentage of how human that work is. And so a lot of my students come out as about 60% human. And so I'm wondering what are what are they doing that these these papers have suddenly gotten better, but they're not using these AI prompts that I said they could doesn't seem like they're using AI. Overall, I'm seeing what is clearly their own language. Sometimes I'm seeing what is clearly their own ideas. 

Eric 21:00

A yet they have improved without the pure AI. 

Dr. Josh Stout 21:03

I've seen a sudden improvement and grammar has, is no longer an issue with most students and and they're coming out when I see the sudden improvement with perfect grammar from someone I wasn't expecting it from, they come out as 60% human. So why are they not 100% human? 

Eric 21:21

Do you have an answer to this? 

Dr. Josh Stout 21:22

I do. I do. I figured out what they're doing because I asked them. 

Eric 21:25

And they told you? 

Dr. Josh Stout 21:27

Well, because I said, Look, I'm not going to give you a hard time. I just need to know what's going on. And every time they would do these weird things, I would ask them, some of them, they lied about the text spinners. I had to figure that out. They're like, No, that's my own paper. I'm like, No, no, it's not no human rights that way. But they, they, they didn't lie about this because, one, I had a really good relationship with some of these students. So they would they would tell me what they were doing. And I said, just tell me I really was. 

Eric 21:59


Dr. Josh Stout 22:00

It's Grammarly. So Grammarly Grammarly is a program that figures out if your grammar is in the right format and students had been using it, but it wasn't helping them much. But starting about a year and a half ago, two years ago, Grammarly hooked up with Open. I and so Grammarly is now rewriting their papers for them. And it's not a prompt. So it's not, it's not what I'm asking for in my syllabi, but they're taking each section of their papers that they write, so they're writing a paper and they're running it through Grammarly and it's fixing everything. 

Eric 22:39

So Grammarly now is both fixing the grammar and doing and I push through it as well just. 

Dr. Josh Stout 22:45

To make sure everything comes out nice and smoothly. But, but it doesn't know anything that the student hasn't given it. So it's still all the students work sort of. And I'm not completely comfortable with this, but I'm growing comfortable because what I'm seeing is students who would have a really hard time with a pure writing class doing much, much better. Because of this, they are able to turn in papers where perfect subject, verb agreement. They they if I say you need to make a topic sentence on this particular topic, they can say, you know, Grammarly, I need a topic sentence that says this Grammarly will fix it. I'm getting a lot of really good papers without the same errors as before, and I feel that what the students are doing is thinking in a useful way about these things and then running it through Grammarly and turning in good papers. Grammarly is not writing the paper for them, but it's fixing the papers. And at first I thought, this is horrendous. They'll never learn anything. 

Eric 23:48

They'll never learn. 

Dr. Josh Stout 23:50

This is what my, my, my teachers told me about spelling, that as soon as I get a spellcheck, I will never learn to spell. For me, this was always going to be true. 

Eric 23:58

You were never going to learn to spell. 

Dr. Josh Stout 23:59

I was never going to learn to spell. This was not a question in my mind and it. 

Eric 24:05


Dr. Josh Stout 24:05

Shouldn't have been in there. 

Eric 24:06

But your students at their age could learn to write English better and it would actually improve their life if they're going to be here. 

Dr. Josh Stout 24:12

Well, what happened to everyone else was spell check is spell check taught them how to spell because they kept underlining things that were the same word every single time. And it would remind them that this is how you need to spell it. And, you know, so when I was in fifth grade, I was held back from every recess because I couldn't spell the word because and I had to write because double the amount of times I say as the last recess, every time I misspelled it, which was, you know, several times a day. 

Eric 24:42

Was this in the U.K. or here? 

Dr. Josh Stout 24:44

This was in the United States. And so I was writing because hundreds and hundreds of times every recess. 

Eric 24:50

I still causing me pain here. 

Dr. Josh Stout 24:52

Every. Yes, exactly. No. So that that that that scene in Harry Potter, where he has to write something in the period, the words appear in blood on his skin that's taken directly from that approach to teaching. It was like it was like carving these words into my own skin. It was it was. 

Eric 25:07

So painful, the word, because now. 

Dr. Josh Stout 25:09

Not really. If I work at it, I can mostly do it, but I'll be typing along and I'll see the little red underline. I'll go back and change it because. Because it's a word that often comes out wrong. 

Eric 25:19

So are you maintaining that maybe people are learning how to write more slowly and differently, but they're still learning how to write? Is that what you're saying? 

Dr. Josh Stout 25:27

Yeah, I'm trying to say that like a lot of people learn how to spell from spell check. I'm seeing in the papers they write for me in class that are handwritten. They're there, their language is better than it had been, and it might be that they'd all just suddenly been taught better. And suddenly I've got a better crop of crop of students. I do not believe this. Every other professor at the university is talking about how terrible the students are, and they're getting worse every year. Sorry, students out there. I shouldn't have said that. Hmm. How do I say that? Better. There are. There are worries that the universe concerns. Yeah, there are concerns that this the students are are are not as good as they had been in previous years. I actually personally do not think that the problem lies completely with the students. I think it lies with the professors as much as it does with the students. As we get older, we get pickier in our lives, and many of the older professors remember how they were taught. They remember hand drawing range maps. They remember cross-hatching pictures. 

Eric 26:34

Basically, you're talking about hazing. 

Dr. Josh Stout 26:37

It's not quite hazing, but there is there is… 

Eric 26:39

Which is deep in the university culture. 

Dr. Josh Stout 26:42

There is a feeling in the university culture that, you know, we have standards and we want our students to meet them, whether those standards are important for the rest of the world or not doesn't always matter to the university. And that's where it comes to hazing. 

Eric 26:56

I like this because. But but your sampling is way, way, way too small. Like. 

Dr. Josh Stout 27:01

Oh, absolutely. 

Eric 27:02

I have been having ongoing conversations about how A.I. is going to make it so that nobody needs to learn how to write anymore. But if what you're telling me is that the right professor can still make students think and teach them how to think and use the tools they have to produce better papers, and then that the the repetition of that pedagogical approach actually improves their writing on its own. Yes, that is a very hopeful thing. But again, your sampling is way too small for me to really hang on to that at this point. Well. 

Dr. Josh Stout 27:35

If a student uses A.I. instead of writing, they won't get better. But if they use A.I. to fix their writing, they will. 

Eric 27:43

Will get better. They use A.I. and the tools they have in order to produce better writing. 

Dr. Josh Stout 27:48

So the reason my students are coming out is 60% human is Grammarly is artificially changing their language, but it's not complete. So it's not a 100% A.I. 

Eric 28:00

Because they did actually write it. 

Dr. Josh Stout 28:02

Right. And so I see their comments, I see their ideas, but I see it written really well. 

Eric 28:07

I am blown away by the fact that these A.I. detectors can actually assess it to that level of of of subtlety and finesse. It's amazing that the way you've described it, it is 60 to 65% their work. They did write it, but then everything else went in Exactly. Readable. 

Dr. Josh Stout 28:28

Right. And so if I if I get a numbers that are if they're not majority human, yeah. I tell them to try again and that they have to rewrite with their own words what what's what's there. I don't do this every for every paper I do spot checks if if a student is is you know, if I see a sudden change in a student's writing, I'll do it. Or if I see writing, that disappears. Too good to be true, I will do it. But, you know, normally I can tell what a student should write like. So you. 

Eric 28:55

Don't. You don't by rote run every student through an AI detector the way the way you run them through safe, assign. 

Dr. Josh Stout 29:01

Write, safe, assign is just a click. Whereas the programs that I use for students to submit their work do not have an AI detector built into them. Yet I think. 

Eric 29:12

I think that's only a matter of time. 

Dr. Josh Stout 29:13

It's only a matter of time. Yeah, it will be built in very shortly. And it's really interesting how some of these detectors work. They they're not just was this copied from a previous paper or does it lack bursting and turns of phrase that are human. They're also looking for the kinds of things that a forensic examination would look at. They' looking at do the fonts change partway through? Was this cut and pasted? Is there line spacing differences? Are the. 

Eric 29:42


Dr. Josh Stout 29:42

Are the editorial choices consistent throughout the paper if someone has a particular kind of so, you know in details, there's something called an m dash versus an N dash. Yes.The  length of a dash in a paper has to be one thing. 

Eric 29:56

To be consistent. Yeah. 

Dr. Josh Stout 29:58

Published papers are all consistent because the editor, the editors will make you do that. 

Eric 30:02

They'll make you use the. 

Dr. Josh Stout 30:03

And so if a student is cutting and pasting from a published paper, there will be a sudden change in editorial choices. Yeah. 

Eric 30:10

You know, the en dash will will turn into em dashes or vice versa. 

Dr. Josh Stout 30:12


Eric 30:13

Like, did you make that choice? Or did that just happen? 

Dr. Josh Stout 30:15

And so these programs pick up on that kind of stuff. Yeah, so so and it gives me report on it as well. And so I then report this to the students and at the very least, it makes their papers better because now, now they're not changing fonts halfway through the paper because they care now and they're starting to do things like highlighting their whole paper and making sure it's in the same, you know, point and the same the same font, which they should have done anyway. Yes. And they should have done, even if they weren't cutting and pasting. But it's the kind of thing that I can now make them pay attention to. So I'm making them better writers on these sort of technical levels, even if that technical level is essentially filing the serial numbers off of their their fake work. And I'm thinking about this. Am I am I just preparing students to not be able to write and get out there and I produce bad papers that they then fix. And I realize if they can do that for most of the time, that's going to be fine if they if they don't write very well, but they know how to put it through Grammarly and turn out a good paper, That's sort of the bottom line. 

Eric 31:17

These tools aren't going away. You know, it's again, I have to keep coming back to to the Sarah Lawrence education. And the question is, can they think?  Like if they can think their way through a problem to a solution, then they can force the tools that they have to get them to show how they did that well enough. 

Dr. Josh Stout 31:33

Yeah. Yeah. 

Eric 31:34

Then they are scientists. 

Dr. Josh Stout 31:36

I can think for you, but you have to be able to set it up properly. And Grammarly is doing very little thinking for you. It's really just rephrasing things into a better format. And so it's still really their own thoughts. And that's why I've decided to accept and. 

Eric 31:51

It's and it's not a text spinner, it doesn't take immune system and turn it into protective framework. 

Dr. Josh Stout 31:58

Exactly. No, no. It gives me good words and it knows when to use the right ones because that's what I does. Right. It predicts the next word in a series of words. 

Eric 32:06

You're really you're really you're presenting a more hopeful integration, a more hopeful view, a more hopeful view of the integration of AI into education and the future of our work then, yeah, I've really seen up till now, well, because. 

Dr. Josh Stout 32:23

I was really afraid and I got an I got it in front of it. I went to the classes that the university was giving me. I studied it on my own. I figured out what I could do. I wrote some papers on my own using AI and they're good papers. They do a great job. Obviously, these are not things I've submitted for anything because they're full of fake data, they're full of fake data and they're fake citations. 

Eric 32:47

That's when you told me that it had made up a perfect experiment and outcome that never happened. 

Dr. Josh Stout 32:52

So I was writing papers where I never did the experiment, and it did the experiment for me, did the results for me, gave me statistics. It was, you know, it was all there and does a great job doing that. So I can be a really useful guide for how to put something together. You could actually get it to tell you how to conduct the experiment you want to do, and it would write the methods for you and then you could do those methods and you would get useful results. 

Eric 33:21

You write it yourself. 

Dr. Josh Stout 33:22

And then you could do it. You could write it yourself. You would cut and paste the methods and anything you changed, you would, you would, you would fix to show what you actually did. And this is this is the world we're going into. There are definitely still fears. So, you know, I know a a young engineer who just graduated from engineering school and he wrote his structural engineering thesis entirely with the AI. And this makes me worried, thinking about structural engineers who didn't actually do the research for structural engineering are out there building our bridges. 

Eric 33:53

Yeah, well, I mean, in some places the structural engineers who did do the work still build buildings. 

Dr. Josh Stout 33:58

I think about that. You know, the guy who just had the workmen hammering away at the pillar that held up the building, they just found that out and he signed off. It said that's not a structural pillar holding up that building, that anyone could have looked at the pillar seeing. Yeah, that's holding over the building is a structure like Jenga with a building. Yes. 

Eric 34:13

Or the, you know, the department store in 1996 when I went to Korea, like at the end of 90 years of, the department store had just fallen down. 

Dr. Josh Stout 34:20

So I suspect that there is human error and there's AI error. And both of these things we cannot get rid of. But I think we'll actually end up in a better world if we can use them in conjunction in AI can be used to evaluate something and say you probably shouldn't knock down that pillar, but you also need people checking the AI because I will hallucinate and come up with wrong thing. 

Eric 34:42

There's no building there. There's just a pillar. 

Dr. Josh Stout 34:44

Don't worry about. Yeah, and really consistently, you know, so there was, there was, there was a new AI that came out recently that was supposed to have stopped hallucination and I asked it a question that I'd asked previous ise and it gave me the same hallucinatory answer as all the other ones. 

Eric 34:58

May I ask what question you're using, or is this a proprietary question? 

Dr. Josh Stout 35:01

No, no, it's a really it's a really interesting one. I wonder if I say it enough if the AI will learn if you read this. Yeah, but I asked it if there is any art carved on two hand axes so hand axes Paleolithic hand axes by definition. And before art begins, right. 

Eric 35:23

Before art had been separated. 

Dr. Josh Stout 35:24

Two or 300,000 year gap between the two. And so by definition there cannot be any art on hand axes. However, hand axes might have pretty stripes on them that the person knowingly included when they made the hand axe, but that was not something they carved in it. It is difficult for I to read Wikipedia discussing these things and understand that hand axes don't have art on it. They see hand axes being made. They define art as something that gets made. Hand axes have an aesthetic value. They define art as having a aesthetic value. And so the guy says, Yes, absolutely. Hand axes are an example of purposeful art showing an artistic result. And that and it gives me examples of it and will say there was carvings on these things, it was decorations on them and say at this particular site, these, these engraved hand axes were found. I know the site that they're talking about. They got it right from Wikipedia. I read the article. There was no engraved hand axes at that site, but there were things that made the I think, this. And so this is the kind of stuff where it's very difficult to know unless you're an expert in the field. So, you know, for example, I, I, I was helping someone make a list of archaeological sites and I said, well, let's just have a I start with the list. And then what we'll do is will go in and check every archaeological site against Google that is listed and see if it really and, you know, did a great job. And we came up with all these fairly well-known sites. But some of the archaeologists who were then checking this could not recognize any of the things that have been found. So even though they're in Google and findable, they're not things that were really in the field of archaeology. So they were still being hallucinated because someone somewhere out there on the web had put something in the popular literature, but it didn't actually exist under that name or context in real archaeological literature. And so AI can't tell the difference between that even when checked by someone who who's checking to see if each of these things exist at that site with that name and and those dates. AI is going to give you wrong answers and lead you astray. 

Eric 37:39

This just goes to this just goes to show there's no such thing as understanding where the AI is concerned. 

Dr. Josh Stout 37:45

AI has no understanding. There's no free lunch. You have to be able to do some of this stuff yourself. And it does worry me as we, you know, there are some downsides. I mean. 

Eric 37:53

You can use this to help you, but you still need to be an expert. You still need to know things. You still need to have knowledge and be able to evaluate yourself. Is this actually better? Is this actually right? 

Dr. Josh Stout 38:03

And a lot of people are going to be using AI as their search. And so AI is going to be summarizing search results and A.I. is going to be what's teaching people in the future in that you search something and find… 

Eric 38:15

But not, if your model is followed along. I mean, you still need a teacher, you still need a person there. 

Dr. Josh Stout 38:20

But AI is not going to tell you that. So if you're a student and you think and you think you can get an answer from AI, you're going to Google it. And AI is going to give you the answer and you're going to cut and paste it. And most of the time you're going to be right. And okay, so here's the other dark side that is coming really rapidly. Then professors are going to take these things and run them through AI to grade the papers. This is absolutely coming soon. When I first started playing with ChatGPT, if you took something and just cut and paste it into a ChatGPT, it would give you a revised summary of it and you could tell you how you want it, how long and what this sort of thing. Now, if you just take a paper and cut it into ChatGPT, it evaluates the paper, it gives you comments and tells you, tells you what how the paper is doing and, what it's doing. It it basically looks at it the way a teacher would. And so I think I think the engineers are subtly trying to bribe particularly high school teachers and saying, you know, you're worried that all the students weren't going to be doing any work. You don't need to do any work. You can now take your students papers, cut and paste them into catchy beat. All the comments will be written for you. You just put a grade at the top, you adjust the comments as you see fit because you know it's your it's your work and it's going to save you so much time. And so this is how AI is really being used. It's being used to save time, and it's being used by people who are intrinsically lazy, which is all of us. And and it's going to mostly be useful and helpful and wonderful and occasionally it's going to mislead people. 

Eric 39:58

Except for the people that it provides catastrophes for because it's misled them and they didn’t catch it. 

Dr. Josh Stout 40:02

Them and they didn't catch it. And this is going to happen over and over again. It's going to happen to professors. It's going to happen to students, lawyers, lawyers. Exactly. And it's going to be happening more and more often. And our AI hopefully will get better and better to try and minimize this. But we're going to have to have I look at A.I., you know, and they'll have to be specialized AIS if you want something for science, you're going to have to have one that's dedicated to that. If you want literature, one that's dedicated to that instead of these sort of general generalists. Once this is where I've been thinking about sort of how AI works a lot and in relation to the human brain is that our brain is made up of something like sections of specialist A.I. material, so that it's these, these, these very difficult to understand feedback loops that even when you know all the components, yo can't really understand what's happening because everything has its own inhibitors on it, on something that's also excitatory and it's all these feedback loops in the brain, each section of the brain talking to other sections of the brain, but also independently running its own sets of calculations. And, you know, so when Freud was trying to talk about an unconscious, he had no concept of how brain actually worked. But now we're starting to see that there are these part parts that are not when you don't consciously understand them, but they're doing all of the calculations that lead us to how we live our life. So something that tells us in our brain is, you know, who is this person? They do facial recognition and says, you know, does this person is this person someone you recognize that's going to be connected to rewards centers? Then I get dopamine. If I see someone I like, I don't get dopamine. If I don't see someone I like. All of these things are prior to our conscious understanding of who I like, this person. And so there's all of these things sort of happening under the hood that then come up to a level of consciousness. Certainly much of philosophy, Buddhism, etc. has thought for a long time that consciousness itself is an illusionary sort of emergent capability built onto all of these underlying aging pieces. And this is this is something that, you know, I think A.I. is kind of working at that. It's trying to create the illusion of all of these pieces at once, which is part of why it doesn't work really well. But it's going to need specialized pieces that work together to work on specific components. And then there's going to need to be an overall AI that is going to have to say, Here, I'm putting it together into this sort of personality. 

Eric 42:30

So you're talking about building a brain from multiple AI’s. 

Dr. Josh Stout 42:31

And building a brain, from multiple AI components, none of which themselves have a soul. 

Eric 42:36

This is an old science fiction concept. 

Dr. Josh Stout 42:38

Absolutely. None of which themselves have a soul. But as an aggregate might have something more than the pieces on their own. So I don't know where you've. 

Eric 42:49

Now you’ve gone a step too far for me, but. 

Dr. Josh Stout 42:51

Well, exactly. But I don't know what a human soul is. I feel like I have one. 

Eric 42:56

All I know is that I feel things and that the machines don't. So that's some that's that's just a bridge, a divide that can't be crossed. 

Dr. Josh Stout 43:04

It's it's it's just one Turing test away and I'll it's it's it's if something until. 

Eric 43:09

They make us believe that they're feeling because they're still machines they don't feel. 

Dr. Josh Stout 43:14

We're a machine. I mean we are incredibly complicated machines but we are machines we are absolutely mechanistic. 

Eric 43:21

I mean dopamine and all the hormones and all, none of that stuff exists in the electronic model. 

Dr. Josh Stout 43:24

No, no. And so this is this is something where I don't know if they'll ever really be able to compete with us in that sense. 

Eric 43:32


Dr. Josh Stout 43:33

But I think that's a that's a computational issue. So by Moore's Law, computers will have as many transistors as a human brain has neurons in the next ten years or so by Moore's Law. So number of neurons in the human brain, number of neurons of of of transistors on a chip. 

Eric 43:57

Do neurons and transistors work the same way, are they? 

Dr. Josh Stout 44:00

Well, here's the thing. 

Eric 44:02

Are they actually comparable in that way? 

Dr. Josh Stout 44:04

In many ways. So a transistor has an input and an output, and then it has a third wire that comes in that tells it whether or not that input, an output should be passed on. Yeah, So, so, so there's a feedback aspect to it and then there's a throughput aspect to it. That is how nerves work. They they have inputs, they have outputs and they have a nerves that come in from the side. 

Eric 44:29

Is it only on or off or is there modulation? 

Dr. Josh Stout 44:31

Well, that's the thing. That's the thing. 

Eric 44:33

Is there gradient? 

Dr. Josh Stout 44:33

Because there are biological systems, everything's working with diffusion gradient. 

Eric 44:37


Dr. Josh Stout 44:39

The, the, the numbers of inputs are, are in the in the hundreds, the numbers of outputs in the hundreds and the things that tell it to be on or off are in the 50 - 60000. So even while we'll have the same number of neurons as as transistors or transistors, as we have neurons very, very soon, we are, I think, many, many, many, many orders of magnitude away from being able to do for computers, to be able to do what people really do. 

Eric 45:08

Right. and that is all affected by all of this gooey, mushy mess that creates feelings like this isn't this never will exist in in no matter how many AI’s you cobble together to create a brain from the Culture series, you're never going to get something that feels anything. It can record inputs. It can report inputs…

Dr. Josh Stout 45:29

See, here's the problem. 

Eric 45:30

It can sense damage, it's not going to hurt. 

Dr. Josh Stout 45:33

We can make it say it does, and we will never know what it really does because all the actual computations are not really accessible. AI’s do not, are not written by people. They write themselves and we really know somewhat like the brain is. 

Eric 45:55

What could possibly go wrong? 

Dr. Josh Stout 45:56

You know, so if you look at any neuron in the brain, it clearly is it's alive, but it's not thinking. It's just sending out chemicals. Yes. The whole brain is doing this. So what you're saying at a certain level is just an emergent property. 

Eric 46:11

It becomes a gestalt. 

Dr. Josh Stout 46:12

Something like that. 

Eric 46:15

I just I guess and this is where and this is where although I am not a religious person, I guess this is where a sense of belief comes in. I believe that I am fundamentally different than any machine that is not biological. 

Dr. Josh Stout 46:27

So so there is there's a wasp out there with with about 12 - 14 neurons, It can dig a burrow, lay eggs, or actually dig a burrow first. Then it goes and gets a spider, knows which species of spider to find. 

Eric 46:44

One species. 

Dr. Josh Stout 46:45

Has a particular array of them. 

Eric 46:47

But out of the bazillions. 

Dr. Josh Stout 46:49

Yes, it knows which ones it wants. Right size, right place goes and finds its burrow again and lays an egg in the spider. And the spider gets horribly eaten by its progeny. This is complex this is something will be very difficult to get a computer to do to to be able to find something in the wild, pick it up, fly it back, inject it in the exact right way with the poison so that the spider doesn't die and stays nicely fresh for the horrible things that are going to eat it alive. All of this is tremendously complex. It is mostly hardwired into the into the wasp, but the wasp has to have a lot of flexibility, doesn't know where that spider is going to be. It has to be able to bring that spider back to its home where, you know, it makes a new home every year or different wasp make different homes. All these things have to be able to be flexible. And this and there's, you know, less than 20 neurons in its brain. So clearly, number of neurons is not the defining characteristic. It's the connections between them. It's diffusion constants relating between, you know, a feeling in some ways is a build up of particular things like dopamine in particular areas of your brain at the same time as, say, serotonin is building up in another part of your brain and your feeling is then averaging between these these diffusion constants that are showing what the concentration is and how far it's spreading in certain areas. These are feelings and they're really, really complex. 

Eric 48:18

And you're talking about the gradient mixture over an actual physical distance of mush. It's it's just no matter how complex you get an A.I. brain to be, you're never going to have that. 

Dr. Josh Stout 48:32

I don't know if that's true. I mean, this is this is what what Turing was asked when he when he when they said, you know. 

Eric 48:37

He was just saying wouldn't make a difference because we wouldn't. No. 

Dr. Josh Stout 48:40

No, no. He he actually said, I think I'm intelligent and I'm a machine. So his his, his that was just straight out reasoning. Machines will eventually be intelligent because people are intelligent and people are machines and end of argument. 

Eric 48:57

But are we machines? Are animals machines, like it's all a machine. 

Dr. Josh Stout 49:01

Yeah. I mean how. 

Eric 49:03

A biological thing that is born as opposed to a silicon thing that is made. 

Dr. Josh Stout 49:08

I don't I don't I don't see how they are different They're just a collection of, of of molecules interacting in particular ways. You can build any one of them. 

Eric 49:17

Okay Dr. Manhattan. 

Dr. Josh Stout 49:18

It's just it's just a code. 

Eric 49:20

You're right. I mean, when you… 

Dr. Josh Stout 49:22

It's just code, it's really incredibly beyond our understanding, complex code. 

Eric 49:27

I can't argue with that. 

Dr. Josh Stout 49:28

You know, I one of my favorite comics xkcd was was we say, you know, people people from the computer world come into biology and say, it's just code. I'm going to solve all your problems right now. And he says, know, look at the optimization for Google. Google's only been around for a couple of decades now the the the the you know code source for Google is is a nightmare of code at this point because eventually it's been building and building and building on itself for decades. We have code that's been building and building on itself for a billion years. So so the optimization processes are really subtle and they're there. 

Eric 50:07

And for those years we've been doing a lot of that optimization very slowly. 

Dr. Josh Stout 50:11

But but involving things like making a new kind of neuron to do a particular new kind of thing and then and then putting those together, right? So, you know, you were just talking about or I was talking about the, the, the ghost in the machine aspect of our brain where we have this consciousness. We think our consciousness sits in our forebrain, the prefrontal lobe. That's what we take out when you do a frontal lobotomy. We think that's where sort of impulse control and all this sort of higher reasoning comes in. Ours is only 66% larger than chimpanzees. Our brain as a whole is about four times larger, so 400% larger. But our frontal lobe is only about 6% larger. We think our frontal lobe does better work than chimps does. You know better by more than just 6%. And it's all about the folding in there. So the shapes of things matter in biology an awful lot. We're only beginning to explore this kind of thing in in in AI and the architecture and how these architectures are going to work with each other. You know, I think most AI's are working calculations at the level of, say, what some of the dopamine circuits might be doing in your in your nucleus accumbens, right? So these are things that tell you whether something is good or bad. 

Eric 51:29

That deep, deep, low understanding, is a basic thing. 

Dr. Josh Stout 51:32

Way, way below most of the, you know, higher portions of the brain, the higher portions of the brain than explain what we what we feel. And so, you know, I think I think I may end up working along those lines again. You know, I don't think it's going to be us anytime soon, but as soon as it can fool us, that's that's what's being said, that it will it will be it will be us. And indistinguishable from and it doesn't really matter if it is actually it's complex. 

Eric 52:00

It's not fooling us yet. And it's as hard as your students work. It still can't necessarily fool you all the time. 

Dr. Josh Stout 52:07

No. And the nice thing is, is is that soul that you're talking about, that that that the bursting is. 

Eric 52:14

I used that word but okay. Yes. 

Dr. Josh Stout 52:16

Yeah, yeah. The thing that makes us different from the machines it's not just the mistakes, it's our idiosyncrasies. It's it's the turns of phrases you use personally. There's a there's an entropy to your language. Yes. And some people have a higher entropy than others. They have a wider vocabulary. They use more variety. Other people have very specific turns of phrase it. So, you know, when Ted Kaczynski was producing his stuff, they were able to identify it partly by looking at the entropy of his language because he'd written a lot. They compare what kinds of words he used to other other other things. I was doing similar kinds of work. And the air detectors are looking for that. And so they're they're looking for, you know, what makes this person human. It's those turns of phrases. It's unusual things that they say. It's relating something to a scientific paper that an I wouldn't mention. So, you know, if I write, if I'm having my students write a paper on, say, blood sugar content, I students will often talk about family members who have diabetes. I is never going to do that. It's never going to just throw that into a paper. My students need me to know these things because they're humans. Talking to a human. And this this is where a personal feeling comes in. And this is the kind of stuff that an echo detector can find and say, No, yeah, that's not an AI. That's a person. They just talked about their mom. You know, these, these, these these are these are things that I would have difficulty for. You know, it's a funny thing in Blade Runner, the voice contest, you know, you're in a desert and you see a tortoise lying on its back in the sun. You know, what's the tortoise? And you know what? Eternal that same thing. You know, that whole thing that was trying to elicit an emotional response in this way. I one of the interesting things of of of Philip K. Dick was he thought that eventually you wouldn't be able to tell the difference. He wondered if we could tell the difference in ourselves are we ourselves AI, this is what some of the whole simulation questions are. Would you even know if you yourself were an AI? 

Eric 54:21

I mean, we're going back to the reboot of Battlestar Galactica, that ends with the merging of… 

Dr. Josh Stout 54:25

And so so with with with, with Blade Runner. “Do Androids Dream of Electric Sheep?” There was a hypothesis that if you let the Androids live long enough, they'd be able to pass the Voight-Kampff test. And what they'd done is they started implanting false memories so that these androids thought that they were human. And these false memories gave them that background. And you could still just about tell who they were with the Void contest. But it took much, much longer. They were able to pass it for longer because they were having these emotional responses that they said that they felt, but they were not real in the sense of a human because they were artificial. 

Eric 55:06

They had never actually had those experiences. 

Dr. Josh Stout 55:07

Never actually had those experiences. Yeah. So I don't know where limit is. I think we're actually really far away from that limit. These are still tools. I don't think we need to be afraid of them. I think in the context of teaching, I think we do need to be afraid of them when it comes to, say, facial recognition by a authoritarian government that's going to be really bad. You combine facial recognition with, say, I of who is likely to act out against the government based on who their associations are. 

Eric 55:35

I mean, we can go a lot further because I can can see I can just track everyone's cell phone and then start predicting where you're going to be. 

Dr. Josh Stout 55:42

Exactly and predict who your associations are likely to be and then arrest you on behalf of these predictions. And, you know, there are a lot of places where a government would be I would think, you know, they're willing to they're absolutely willing to arrest a few innocent people if it just stops any possibility of rebellion, that that wouldn't bother them at all. And so this the kind of stuff where I'm afraid of AI, that kind of prediction. I'm also afraid of AI when it comes to predicting my marketing choices, where it will predict what I want, what I want, what I want. And there's a certain sort of subtle moment where it's telling me what I want, and if it's always right, it knows what I want, and then it suggests something to me that I might want. Now it's starting to make me want this. 

Eric 56:25

And this is what frightens me about, about the coming use of teachers using, and professors using these these tools to grade papers. Because if you begin to you begin. We're now giving up. You're saying that there's a way to get a hold of this. But then if teachers start using it to grade the papers that the students have used to write it, we're giving it we're giving it away. 

Dr. Josh Stout 56:47

And as as our papers move more and more towards an AI optimum standard, English, the AI will pick up on this and refine the standard English even more. 

Eric 56:57

We're all going to be writing and talking like newscasters on the nightly news. 

Dr. Josh Stout 57:01

Exactly. Yes. Yes. That's that's where we are definitely headed, because that's where where A.I. will take us, because it doesn't it's not making these decisions. We are. We are we're we're saying a paper that doesn't have a lot of weird stuff in it is better than a paper that does. 

Eric 57:17

I have to believe the humanity will persist, that we will keep our human nature one way or another. 

Dr. Josh Stout 57:24

No, I'm sure a high school professor will have a paper that gets graded and then at the bottom say, Hey, I saw you at the school Christmas party. Was it was great. It was great hearing you saying, you know, something like that right at the bottom to show that they're human, but the rest of it will all be you know, I think that's the kind of world where we're heading toward where workers are using A.I. to do rough drafts of basically all of their work. Then they're quickly rewriting it and submitting it to a manager who puts it through an AI to see if it's there. The workers will be hired by RDR. The A.I. checks the resume and go through tens of thousands to pick the right person. If I'm writing a resume, I'm going to say, Hey, A.I., what? What words should I put in this resume to get it noticed by an A.I.? And it will tell me. Yep. And, you know, optimization. So I should probably take my podcasts and put them through, you know, search optimization, A.I. that I know exactly what to do there. There are people who are paid to come up with search optimization terms that doesn't need to be done. It could be done by A.I.. Absolutely. I would bet if there are people being paid to come up with search optimization terms, they're already using A.I. to come up with those terms for you. 

Eric 58:35

Well, I mean, we use AI to create the transcript of this. 

Dr. Josh Stout 58:40

It does a terrible job. By the way. I've been reading the transcripts. 

Eric 58:42

I do. I do. I do a little spot checking, to be honest, but I don't go through it entirely. But part of the problem is that it doesn't know when one of us stops and the other one starts, but…

Dr. Josh Stout 58:52

Combined with my sort of repetitive study stuttering and mis stepping on words all the time, it has a lot of problems. And then and then my vocabulary, you know, it did not know who Duns Scotus was the other day when I mentioned it. 

Eric 59:07

Yeah, well, I mean, I try to go through and catch the important terms so that they're so that they are on the web correctly. But yes, I will miss some too. 

Dr. Josh Stout 59:16

But yeah, Anyway, so this for me was was was really hopeful as a as a professor to see my students actually getting better at writing through the way they're using AI, not the way I expected it. I thought that they'd be making their outlines with AI, but students don't make outlines. They just start writing the paper and then something goes wrong, so they fix it using Grammarly. And so it it's not a process. 

Eric 59:39

I hope you all are using outlines. It makes everything easier and helps you write the paper and rewrite the paper. 

Dr. Josh Stout 59:45

Absolutely. But if… 

Eric 59:47

It’s so important. 

Dr. Josh Stout 59:48

At the level I'm working with, where the students are, are reading one scientific journal article and then they're writing a two page single spaced paper on that scientific journal article. An outline doesn't help you much because the paper is already broken up into sections. Those sections already have a particular form. 

Eric 1:00:06

I guess, again, I'm product of a Sarah Lawrence education. 

Dr. Josh Stout 1:00:09

You're thinking of a 20 page paper where you've read four books and you need to think about how those books relate to each other. 

Eric 1:00:15

Yes, I am. 

Dr. Josh Stout 1:00:15

That's that's the humanities. It's a very different, different kind of world. Now, if someone's writing a thesis, I suggest you start off by your your dissertation thesis. Ask the A.I. what to do ask the AI about each chapter. It will make your life much easier. 

Eric 1:00:30

Yeah, it’ll help you write the outline. 

Dr. Josh Stout 1:00:32

Yeah. No, this is. This is what I've been doing, and I absolutely use it that way. And I hope to use it even more. This. There's an awful lot of writing that is not really necessary. That is there to keep bureaucracies alive. Bureaucracies, paper. And so the AI will now be producing this paper and very soon the A.I. will be doing the bureaucracy side of things too. So AI will be feeding itself paper and we can just sit back and watch it all happen. 

Eric 1:00:57

And watch it hallucinate and mess with our lives. 

Dr. Josh Stout 1:01:00

Yes, well, they'll be. They'll be it'll be a while, but eventually a ten year rule will be, hey, AI conducted you will submit your stuff and it'll be come up with an answer. 

Eric 1:01:10

This has been a pretty far ranging conversation. But I have to say, I’m, …  it's really the first seriously hopeful view I've had of how AI can be integrated into a university setting and then positive outcomes can come. So thank you for that. 

Dr. Josh Stout 1:01:26

Thank you. And I think they are actually are as I said, I think my students are actually writing better now than they have previous years. 

Eric 1:01:34

That's that's an amazing thing to hear. All right. Well, thank you, Josh, until next time. 

Dr. Josh Stout 1:01:39

Thank you. 

Eric 1:01:41

Keep writing, everyone. 

Artificial intelligence - Wikipedia

Grammarly - Wikipedia
Grammarly: Free AI Writing Assistance
Grammarly makes AI writing convenient. Work smarter with personalized AI guidance and text generation on any app or website.

Article spinning - Wikipedia
Free AI Article Spinner
Free Article Spinner is an AI-powered tool that rewrites and transforms content into engaging articles, removes plagiarism, and offers unlimited spins. Try it now.

SafeAssign | Teaching and Learning with Blackboard | Northern Illinois University

ChatGPT - Wikipedia

Turing test - Wikipedia

Blade Runner - Wikipedia
Voight-Kampff test
The Voight-Kampff test was a test used by the LAPD’s Blade Runners to assist in determining whether or not an individual was a replicant. The machine used in the test measured bodily functions such as respiration, heart rate, blushing and pupillary dilation in response to emotionally provocative questions. It typically took twenty to thirty cross-referenced questions to detect a Nexus-6 replicant. The test was utilized by the LAPD Blade Runner unit from its formation in 2009. The department’s fi
The Science Behind “Blade Runner”’s Voight-Kampff Test
Is Rick Deckard a replicant, an advanced bioengineered being? The jury concerning the character in 1982’s Blade Runner is still out. Harrison Ford, who plays Deckard in the film, thinks he’s human. Ridley Scott, the film’s director, is adamant that he’s not.* Hampton Fancher, the screenwriter for the original film and the sequel, Blade Runner […]
Quiz: Are You A Replicant?
Take the Voight-Kampff test that Rick Deckard and all Blade Runners swear by.

Consciousness - Wikipedia

Theme Music

Theme music by
sirobosi frawstakwa
sirobosi frawstakwa | Instagram, Facebook | Linktree
View frawstakwa’s Linktree. Listen to their music on YouTube, Spotify here.