Today's eerily impressive artificial intelligence writing tools present us with a crucial challenge as writers do we unthinkingly adopt AI's time-saving advantagess or do we stop to weigh what we gain and what we lose when heeding it's siren call . Today's guest, a linguist and educator leads us on a journey connecting the dots between human literacy and today's technology. It's tempting to take the easy way out and let AI do the work for us.
Our guest cautions that this efficiency is not always in our interest as AI piles with suggestions or full blown text we risk losing not just our technical skills but the power of writing as a springboard for professional reflection and unique expression. It is an honor to welcome professor Emerita of linguistics in the department of world languages and cultures at american university in washington dc, author of ten books and another one on the way, and the focus of today's episode is her book.
"Who wrote this: how a i and the lure of efficiency threaten human writing", Naomi S. Baron welcome to the show so much, Aiden. I'm delighted to be here. . You point of all this in the first place and i thought i'd start with something a little bit off the beaten track here. i was gonna do this as the intro but i thought people would actually go, Am I on the right show?
So I fed Google's Notebook LLM with a number of your YouTube interviews, and I got it to generate its own interview about that. So I'm going to play a little bit of that, and it will inspire us to have a little bit of a chat about what came out of this and indeed your own experiences with this. Okay. eye. Amazing, talk about a long view. Seriously. Yeah, so we're gonna try to unpack, you know, what she has to say about AI changing how we communicate like The whole landscape of language.
I'm excited. This could be good. Me too. And honestly, what strikes me is how professor Baron gets at this core question. Like what even is language, right? She digs into research from way back, like teaching chimps sign language. Really? Yeah. And how all that helps us understand, you know, AI and language today. It's really fascinating. Yeah. Cause it's not just about AI spitting out words. It's about meaning, you know, intent, even like emotion. Maybe. Totally. And that's enough of that.
Now let's get back to the humans. You had your own experience. It's a really good, it's absolutely incredibly good. A Dean of a college I know said podcast. Now these guys are better than you. They don't make the same mistakes you make. I think people like my mistakes. I heard about Deep Dive, Deep Dive being the audio option in Google Notebook LM. This audio option just came online in September, I believe, mid September.
And there was a whole rash of articles and, the New York Times, and the Washington Post, and the MIT Technology Review, and so forth, saying, wow, do you know what this is? So I thought, well, let's find out what it can do.
And what I did is I took a chapter that a colleague and I had written for a book that's just about to come out, and then I, and I put that into Notebook and said, please generate me, I just pushed the button, that was my way of saying please to generate this roughly 10 minute long deep dive that creates a podcast with a he and a she voice based on the material that's in that in that notebook, namely my chapter or my book proposal.
The, when I first listened to the book proposal, I literally had tears in my eyes. Not because I loved hearing about my work, but because of the incredible ability of AI to do things that a year ago. probably even six months ago, we could not have imagined possible. Not just the quality of the voices, and, and, and the she voice is way too chirpy and eventually grates on you, eventually meaning after Totally.
Totally Naomi. 15 seconds and I've had various people listen to this comment and I can't listen to this anymore. The content for the book chapter was actually pretty accurate. And it was eerily Then came the book proposal. And actually, having just sent the written proposal to my editor I sent the link to the deep dive. And I said, So now you don't have to read proposals anymore. All you have to do is listen to them.
And she immediately got back to me and said, But what's totally missing, totally, from this podcast is your humor, the nuance, the literary examples that you included. The asides. The, the character of your writing is totally missing. So no, I'm not going to use this for reading proposals because I'd have no idea how this author could write. The other thing that happened is that the first half of the podcast was accurate in terms of the content that I had in my proposal.
But then it started veering into things that might have been nice for me to do but were not part of my work. And it reminded me of what's happening with the company Perplexity now.
Which is being sued by a number of news outlets and what Perplexity is doing is taking as training data, a lot of news information, which it may not legally have the right to, the courts will decide, and generating So that's what's, is accurately in the original news source, and then spinning its own information that's not in the news source, and that's not accurate.
So, anyone who tells you, well, if you feed in the data to an AI, and only the data that you want discussed, in whether it's a written form or whether it's an audio version, because the audio overlay these days is comparatively simple. And we're not going to have any hallucination of the sort of stuff that's false. We're not going to have any bias and all that kind of stuff. Do not believe it, because I heard it with my own ears.
It brings up something and I have to apologize to our audience as well. I am definitely not going to be AIable in today's show because I'm going to go all over the place. And I'm just going to spark off things that you say, Naomi, because. One of the things there is authorship. So if I, for example, played a clip from a movie in this show, or I played some music that I didn't have the rights to what happens on YouTube is you can't have any monetary gain from it.
So it's, exempt from having any type of advertising and it's flagged to the owner of that. And when i play that piece from the lm from notebook i was like on who owns that and i know you talk a lot about this because this is something that's in the minds of many people maybe somebody has, used an ai heavily for their book maybe somebody has written music with it maybe somebody has created music with it or visuals and oftentimes they'll think is like who this.
is the fight that at least in the United States is aggressively being waged right now. Perhaps the first lawsuit more than a year ago now, came from the New York Times, when it said to OpenAI and to Microsoft, thou shalt not have this data unless you pay for it. And that's, Case has not been settled, but other kinds of things are now happening that are terribly worrisome. It's not just the news outlets that are, at least in the States, signing licensing agreements with a number of the companies.
So let's start with OpenAI and, and with Microsoft and with Google and with Meta. But it's also publishers of. books, and publishers of journal articles. So, one of the most recent things that happened is Taylor Francis, which is a huge publishing conglomerate. It publishes lots of academic journals. I have articles in Taylor Francis journals. I have a book in a Taylor Francis company, Rutledge.
And that licensing agreement was signed by Taylor Francis with whichever of the large big tech companies it signed with. And the licensing monies go to Taylor Francis. Now, for journal articles, Taylor Francis, like most academic journal publishers, Taylor Francis holds the copyright. You know, you're not allowed to hold, have copyright for most journal articles, academic journal articles. But for books, thank you, I hold the copyright, which means in principle I should be getting some revenue.
I mean, five cents, you know, two pence, , you know, a shilling. I, I, it takes whichever level of, a euro, a penny, but whatever you want. It's not for the money. It's for the fact that it's, I hold the copyright, but I have no control over how my copyrighted material is being used. And that is extremely troublesome. In some ways, this is reminiscent of what happened when Google Books started scanning books. In order to make information free, right?
And then there was the question of how much of the book is exposed on Google Books. And some pages were available, some pages weren't. Some authors allowed all of their books to be there. And then a number of organizations in the United States the Authors Guild and I think it was the American Association of Publishers, sued and said, wait a second, you are taking copyrighted material and making that available. For a decade, the lawsuits went back and forth, and eventually they would be settled.
And then they were reopened, and it took 10 years to finally say, I guess this is fair use, Google, you may post what you want, as long as it's snippets, but snippets can be huge amounts of text. So, it's this, it wasn't for the money that the authors wanted to have protection of their copyright. It was because they were, They were the authors, and they wanted rights to control what was done with their work. And they lost it with Google Books.
And I don't know what's going to happen, nobody knows what's going to happen, with the suits that are going to come, either from individual authors or, you know, Groups of authors who are having their style stolen from them because you take all the books of name your favorite author, living because it's easier to sue if you're alive. And eventually the copyright does run out, both in the UK and in Europe and in the United States.
But before then, if you have a live author whose work is being taken, there are many lawsuits currently going on saying you know, to the likes of OpenAI, you may not take my work. And OpenAI is saying, I didn't take your work. All we did is took words, and we clipped some words here. But we also know your writing style. No, we're not talking about that in the lawsuit. And we can create something in your style. There was a case where the writer Gay Talese, had a number of his pieces.
It was a journalist for the Washington Post. It was a legitimate experiment. Took the work and fed it into an earlier version of an AI and generated some new Gay Talese text. And when played to Gay Talese or shown to him because it was written text Talese said it's not me, but it's Not bad. It's sort of close. And now in the end of 2024 it's going to be darn close. And the only thing missing is the particular plot or storyline.
That thing about it's good enough this worries me a lot i was telling you before we came on air on the lecturer and i'm just coming up to lecture season and for years i mean i started this module. Eight years ago and i've always had AI in it and how i was gonna rewrite the rules of society and you need to be more adaptive you need to have more resilience because you will lose roles very very quickly and one time a student.
As a joke, made a tinfoil hat and left it on my lecture and my lecture, which was brilliant. I thought that was great creativity with a big C , but my point was always that this was coming and that I think they only started listening last year.
The last year was the big shift but one of the things i was telling them always to use was tool so a tool came out called essay bot and it's still available still really really good and i say but actually has a button that goes deep plagiarize, warning comes up and goes don't worry if this is going through a plagiarism checker it won't be detected if you press this button right so i was telling them this and the thing was they thought it was like entrapment, I will know if it's written
by the AI, but today I don't know today. I checked it today before we came on air to see how good and how advanced it has become. It's amazing. And it actually comes up with stuff.
What I like using AI for is to come up with stuff you just didn't know about, or some, Piece of information that you just didn't think of but i'm saying all this to say and this is how i found your work and i'm so sorry i didn't discover your work earlier but i was writing an article and i was writing about the fact that part of the gift of writing or creativity is the struggle it's the, The.
You have to go through as Yuval Noah Harari would say that the desert of boredom that sometimes is the case it's like for me it's like going to the gym going to the gym, you do the work you show up you just keep doing it eventually something will happen for the better you'll get injured or you'll get in good shape but the same for for writing for me and i say all that to say.
That's how i discovered you cause i was like surely if you're not exercising this muscle this mental muscle i know about digital dementia but i was like i'm surely you're gonna have some type of atrophy in the brain and this is, at the heart of your work the understanding of this and i'd absolutely love you to share this because if anybody gets anything out of the today show.
Please listen to this because this is the point you will lose it if you don't use it you will lose it You have given me about an hour and a half worth of things to respond to, so I'll take them in sort of the order if I get it right, that she's that she's going to raised issues. Tools that can spot plagiarism.
Well, GPT 0 was created about a month after, Edward Cheng, I think created it about a month after Chachabiti came out just saying, I wonder if there's a way to figure out if something was written by a human or not. And that tool has gotten incredibly better over the last two years.
I'm impressed with the work that he's doing, and he's doing other kinds of things that are helping people become better writers themselves rather than just a tool for faculty to say, gotcha, I figured out you did this. But almost every text generation tool has some kind of plagiarism spotter in it, and whether they're accurate or not, who knows. In the United States, before ChatGPT came out, probably the most popular text generation tool was called Pseudowrite and it was using GPT 3.
That is, what OpenAI has done is licensed first it licensed GPT 3, then GPT 3. 5, and now it's licensing GPT 4. Will it license GPT 4? Oh, I don't know. We'll find out. But there are lots and lots of companies that are using the technology that OpenAI has available, and were doing so before ChatGPT came on the scene.
So, grammarly, a tool used heavily in the United States, not just by students, but by lots of people in the workplace, and used around the world by people who are writing in English, particularly people who are not native speakers or fluent speakers of English are using it. And it has had a plagiarism tool within it. So, will plagiarism, Always be spotted?
No. And the reason for that is the way that AI generation works is sometimes it will take whole sentences rather than just individual words when it goes to predict what the next most likely word is going to be, which is the way that LLMs work. And sometimes it's more word salad and you'd have an incredibly difficult time saying, Oh, this came from so and so. If you're a really good writer, and reader of literature, you may be able to sense the voice of a writer.
But if the prompt has not been right in the voice of so and so, you know, make it sound like Charles Dickens. Make it sound like Shakespeare. Make it sound like Because you can do that. and do a fairly good job, but if you haven't said, make it sound like so and so, then probably you won't be able to tell the author's voice. Okay, so what does this mean going forward? You talked about mental muscle.
Mental muscle was a theory put forth in the, well in the 19th century in particular, and I can speak for the United States, I can't speak for other countries but it was a theory put forth as to why it was really important to learn Latin. And some Greek would be a good idea, too. Now, were you going to be, in your daily life, after you graduated from college, were you going to be using, were you going to be writing Latin orations?
No. Did you really need to suffer through memorizing all those conjugations and declensions and reading Caesar's Gallic Wars and so forth? Did you have to do that? The rationale that was given is that it built up your mental muscle to be used in other kinds of activity. So it was the journey that was important, not That which you were using that journey for while you were in school. So, as you come to writing, the question is, what's the relationship between writing and thinking?
I'll get to the atrophy in a moment. But first let's talk about the relationship between writing and thinking. If I were to take both hands, and both feet, and we use the fingers and toes. I would still not have enough digits for citing the, the writers who, for at least three centuries, have said there's a connection between writing and thinking. You know, there are many versions of this. Flannery O'Connor is the one most often cited in the United States. Joan Didion is often cited.
And what is it they're cited as saying? I only know what I think. When I see, what I say, and say, meaning here, what I write. I only know what I think when I see what I write. Or I only can think when I write. And what they're telling us is that the human mind is a miraculous, I mean they're not saying this, but they're presupposing it, is a miraculous organ, but one of the things the development of literacy did is it made possible New ways of thinking, and people have written about this.
Eric Havlock wrote a book, he was a classicist wrote a book called Preface to Plato in which he argued, and I agree with half of what he says he argued that the development of the Greek alphabet, which was the first alphabet where you had one symbol for each sound, and it was derived from the Phoenician, Phoenician being a Semitic language, Greek being a, an Indo European language, so they work differently in terms of their sound system. The Semitic system has a so called consonantal alphabet.
We only represent the consonants with a symbol, and you sort of have to know when you're reading. which vowels are going to be. And this is still the case with something like Hebrew. But what happened with Greek is you actually developed, it was around 800 BC individual symbols for each sound. And the argument that Havelock made is that, that it was the alphabet that made possible Greek philosophy. Well, I don't think, I don't agree with the alphabet part, but the writing part for Darnshore.
There are other things going on in archaic and then classical Greece that made for the, the Plato's and the Aristotle's of the time. Okay, but what we do know is if you can write something down, you can look at it and reconsider. You can say, is this really what I mean? Is this what I think? It's interesting when I was writing Who Wrote This?, and I was working on the chapter on creativity. I didn't know what I was going to conclude. I knew the research that I wanted to go through.
How I wanted to begin the chapter with talking about the birthplace of Leonardo da Vinci and how we, we honor him as much as we do his work and so forth. But I didn't know what I was going to conclude until I was about three quarters of the way through the chapter and I said, now I get it. And what is it that I got? That human beings care about human beings being creative. They care about their own creativity levels, which may be fairly minor. Or they could be pretty good.
You could win the Booker Prize for your book. You could win a Nobel Prize for your book. But we do smaller things as well. You could get a a blue ribbon at a a state fair for your apple pie. State fairs are big things and county fairs are big things in the United States, but there's still a tad of agriculture left. So you feel good about that. You feel you have created something. Does the world care? Well, not particularly, but creativity functions at many levels.
And the conclusion that I came to in the chapter on creativity was that I don't care. how good AI is at creating art, at creating music, and at creating writing. What I care about is whether human beings can create, and what it means to us as individuals and as societies to have created something. The example I, I was giving a lecture, I was part of a panel discussion a couple of days ago, and the example that I gave was this.
You walk into a museum, and I was talking about the National Gallery here in Washington, D. C., and I see, you see it all the time. People walk up to a painting, they see, Aha! This was a Leonardo da Vinci. And then they step back and they look at the painting. But they only were interested in the painting because they saw who created it.
Now you may say, This is not a good way to appreciate art, but it's a good way to be part of a society that has said, for right or for wrong, this is a great artist. When the Impressionists started out, and they were given this nasty name, Impressionists, this is not actual painting. You know, Van Gogh's stuff is junk. I mean, I don't even see a real cypress tree there, right? But we changed our minds over time.
And now we come up, you know, there's a big Impressionist exhibit going on right now in the National Gallery. And I promise you, there are people going up and reading the tag, and then standing back and saying, Yeah, that's a really impressive, I'm impressed with the Impressions painting. So why do we do that? Because creators are part of our culture. And AI is not, and never will be, part of that human culture.
beautiful absolutely beautiful so i was thinking about this that a couple of things that the first is that idea of, value attribution you only give a value when you know who created it. That's a we see that on this shows the innovation show we see that in innovation all the time it's. It's not an idea meritocracy it depends on where the idea comes from which actually should have no impact on the idea the idea should be the idea that's one thing but the thing that i really.
I set out on with all this odyssey of your work it was. I remember when my older son when he was younger was terrible at sport like i was as a kid i was often last picked he was absolutely that kid and they did a school sports day, i need the race so that is big long race around these football pitches a few times and he was so stressed about doing it and i just said just finish it.
That's all it doesn't matter if you realize just finish so he did he finished and loads of kids gave up when i was and i went go to work so i left i waited till he was over went to work, came back later on and there was a medal at the door just on the ground and people think dad i came in and said what's the medal for and he's like oh they gave us a Really proud that you finished but you didn't win and it was one of these moments that was i think i think it was so important i hope it was
so i said to him, that's not how the world works you don't get a medal just for being there and he didn't care about the metal that was the point he didn't care about and i was like oh maybe we should just throw the metal away and he threw it away right so i bear with me here, So years later today, he's very good at jiu jitsu and MMA and he won his first gold medal a few years ago. He treasures that medal and I treasure that medal because I know how much effort went into that.
Years of toil and effort and failure went into that medal. And I say all that to say , for me, when you're, when you're a piece of art, particularly in a piece of art, like in the Leonardo days, the struggle of the artist to create that with the tools they had in place is incredible. And I think that's what you're valuing is this struggle.
And that's where this all came from for me is that if you deprive a child or a student the journey of the struggle they don't build that mental pathway that neural pathway, i'm there for the future of humanity is gonna be making poor decisions if they can make decisions at all and it's that piece that i thought was so fascinating about your history not just this book but the echo, into the past of all your other work and how that's all accumulated to give you
this unique lens to comment on this. All right, but look, There are a couple of comments, and some will agree with you for throwing away the medal, and some will not. All right. I didn't, by the way, I didn't throw the medal. I, I may have hinted that it should be thrown away It doesn't matter whether it physically was thrown away or not. All right, so one of the problems in the United States right now, is grade inflation.
So, you have a child who's seven or eight years old, who's getting all A's. Well, they shouldn't be getting grades at that age, but they are. You have colleges and universities, the kind that I teach in, where the average grade, once upon a time, was a C. You know, A is best and E is not so great. F is And the average grade at my university now is an A There's something wrong with this, and this is typical across the United States.
So it's sending a weird kind of message to students that you're just doing well, and there are reasons for the grade inflation that I won't go into at the moment unless you want. Similarly, if you go um, and there's some kind of competition, you know, you're having an archery day, or you're having a swimming whatever, and everybody gets. A medal, a something.
I remember my son was in a camp, and the phrase used by the athletics counselor at the end of the competition was, We're all winners at And the name of the camp. Now, interestingly, it was a music camp. And they didn't do this for music. So if you wanted to play in a solo recital, you had to work really hard. And you had to get selected. And there were different levels of orchestra. And if you weren't that good, you were in the middle level orchestra.
And if you were really good, you had had to work at it. Okay, there are, uh, are ways of telling people we're all good, and did my son think he was a winner? No, he was sort of like your son, in that he wasn't great in athletics then, but he has been since then, including in MMA. Okay, but then there's the question of, does everything that is creative require struggle.
I think of some of the things that I write, and some of them are really hard, and they're a struggle, and I end up doing seven, eight, nine drafts before I'm even quasi happy with them. And other times, even writing the same sorts of things, I just sit down and write.
And, you know, I clean up the language, and I get rid of repeated vocabulary words, and I try to make it more friendly, because, you know, first time out of the gate, my writing tends to be very formal, and I've come to write more informally over time. But in terms of the idea and the structure, sometimes it just comes. And that is true, I think, you'll find with some other writers. It's true of some painters. In terms um, you know, take classical music.
And I'll talk specifically about classical. You look at how much music Bach wrote, and I promise you, he didn't struggle over every piece. He would have been long dead had that happened. And you look at Mozart, and Mozart died young, but he has a huge oeuvre, okay? So, you know, Was he a fantastic composer? Yes. Was there struggle in every piece? No. There couldn't, I mean, he had his own struggles. He didn't have enough money.
He had kids he couldn't feed them, you know, and they had health issues. Sometimes you just do your thing. So, do you have to learn at some point? Surely. Take Picasso for a moment. Interestingly, when you look at Picasso's early work, and Picasso talked about this a little bit as well, it's very classically trained. He learned how to do art. And then he decided, that's not how I want to do art.
And I suspect that a lot of his later works didn't take a fraction of the time that his earlier works did, because he had become himself, okay? But the new things that he was doing weren't as using the techniques of the classical style that he had spent, I'm sure, years perfecting. So what does it mean to say it's the struggle for the newer Picasso? I don't know. He has since died, so we can't ask Um, But I think it's, it's a more nuanced story. Developing a craft is critically important.
You're not going to get a gold medal in in mixed martial arts, unless you have trained your body. And your body has changed, and it's become a new you, as it were. But, you know, Grandma Moses did paintings, and she didn't have a whole lot of training, and her paintings sell for a great deal, thank you.
So, it's, it's, It's a more complex story, I think, and it's going to vary by individuals, and even for those individuals, it's going to vary by the time in their life, and whether they just feel, I can sit and write this piece.
I mean, I have written op eds before, which means you get about 700 to 900 words, and re written, and re written, and re written, and re written, and then sometimes I'll just write it, read it twice, and send it off, and the editor says, gee, love Um, And the thing that I spent so time on, uh, I wanted to share this with you because a guest on the show shared me this brilliant quote, a guy called Peter Compo was on the show. His book is here at the emergent approach to strategy.
And he was talking about. innovation and he was saying how innovation is a highly disciplined skill while people think it's messy you know it can be messy but it starts with a discipline so what's the challenge i'm trying to solve trying to what's the problem what's the bottleneck i'm trying to get through, and he shared a quote by charlie parker and it was like master your music master your instrument then forget all of that and just play, Yeah.
Yeah. that you can have moments where you get into the flow state and you nail it but it comes from all the work all the reading all the struggle of that that's that's what i'm trying to get out that.
You mentioned in the book, the Norwegian system, when ChatGPT came out, there was a huge outcry against the government that this is going to destroy education and that would be one of my concerns for society is that if students do this more and more and they outsource the thinking, the reading, the, the YouTube video, like we did with the LLM for your interviews that's parsing of the knowledge is not me doing it and yes i can be absolutely
a pain in the butt, what i need to do it in order to be able to let this marinate and it's the marination like when you write those Op Eds, your Thoughts marinate for a long time. Fusing, maybe discussing with other people, trusted advisors, partners, and therefore you kind of hone, you chisel away that idea and you eventually come out with something. And if you don't do that at all, you miss all that. And that, that's where I'm saying, that's the gym work.
That's where you're building the but let me make a little distinction here. For some things, I have been thinking about the topic for a long time. When ChatGPT came online on the 30th of November, 2022, I started writing A piece that eventually ran in, the Conversation, in January. Now, I had not been thinking about ChatGPT, because ChatGPT did not exist before then.
I'd been thinking about thinking, I'd been thinking about writing, but here was something new that I was writing about, and I had to ask, how is it potentially going to have an impact on our writing and thinking? And I hadn't thought about those things. Until I sat down to write this piece saying, I know it's going to have an impact. Let me try to figure out what it might be. So, I had written lots of things over the years, but not on this particular topic.
Alright, so talk about use it or lose it skills and whether that is , a foreign language. Which is one of the areas where it's pretty clear, if you don't keep honing those, keep practicing those skills, you're probably going to lose them. , , We know that there are lots of AI tools being used by law firms now. Increasingly, they can write, the briefs for you.
They can pull cases, and there have been some famous examples in the last two years of lawyers using ChatGPT to pull cases, but actually those cases didn't exist and it was the judge who figured out they didn't exist, and these are great examples. But if you put into your LLM only real cases, and you can do that, and that's been happening actually for several decades. There have been ways of putting them into, it wasn't an LLM before, it was just a database. So you know the cases are real.
you have your brief with all the cases cited automatically generated. Do lawyers remember how to do this anymore? There are law firms that are saying, Well, we have to let our lawyers practice every once in a while, doing it the old fashioned way. Because otherwise, they're not going to know how to do it. I probably put in the book, because I cite it I can, uh, myself, saying to my students, What do you know when the internet is down? What do you know? Can you write by hand? Can you spell?
What do you know? And I would ask the same question. What do you know when OpenAI is not charging in US dollars, 20 a month for the latest version , of GPT, whatever it's going to be? Or you didn't have that money, or they upped the price, or whatever it is, and you can't use that tool, and, and Meta doesn't have Llama for free, and whatever. What are you able to write? And I'm not even asking, are you going to get a job, compared with someone who has the skills on how to use this.
What's in your head? An example that I'm very fond of that is not mine, but it is from something about 30 miles down the road from me, namely the U. S. Naval Academy in Annapolis, Maryland. The Naval Academy decided a couple of years ago that it was going to train the students in the Naval Academy how use, , Old fashioned navigation equipment. And the reason they did that was not because they were worried the satellites would go down or get because there's a lot of fear of cybercrime.
They wanted the midshipmen, as they became officers, to be able to navigate At C, if they had to, the old fashioned way. So, the same is true of why I believe people need to write themselves. I don't care how good Chachi BT gets, or the other equivalent LLMs. I don't care. I care who we become. So this is, in essence, a version of the creativity argument. I care that we have the skills. I care that we have the will.
I care that we're willing to, and this goes to your put forth the effort, that we're willing to put forth the effort to do writing that represents who we are. So part of the question is, do we care that the writing is an expression of us?
I wrote a lot in, who wrote this, about the importance of personal voice, of whether that which Comes with your name, attached to it has an essence of who you are of, of what your style is, of, of of what's in your head, of what anecdotes you have, because you have this lived experience. When you interview students about how they feel about using these kinds of tools.
A larger number than you might predict will say things like, I don't want to use these because I don't want my writing to sound, and I'm quoting , a survey that was done in the United States a couple ago, um, of what do you, do you use these tools, do you like these AI tools and so forth, I don't want to sound like a box of mechanical pigeons. I want the research, I want the ideas, I want the sentiment to be mine.
And in other kinds of work I've done, and there was a little bit of this represented , in, in, who wrote this, on handwriting versus , on a digital I surveyed students in the United States and in Europe, heavily in Italy, but other European countries crept in , And What the university age students told me, over and again, was when I write by hand it's me writing, it's real, it's an expression of who I am, I think better, I concentrate better, , and all those kinds
of positive accolades, simply by writing. By looking at the physical connection you have with a writing implement and some kind of, we'll call it paper, because that's mostly what we write on by people know the difference, they notice the difference, they feel , it's an encapsulation of them, whereas putting fingers on a keyboard overwhelmingly, And if you think about what AI generated text is like, it's your fingers on a keyboard putting in that prompt.
Oh, these days you could also use voice, I know that. But everything is digital, and there's no, there's no reality to it. There's a move in psychology, embodied cognition, that says to know. And to use your brain to come to know, or to come to think, or to come to feel, is not just something that's happening in your cerebral cortex. It's happening in your whole body. So I see the way that you're sitting right now with your, you know, just sort of, I don't have the fingers quite right.
Okay, so there's this thinking of, the feeling that, I'm assuming that, you know, maybe I can think about what she's saying more if I, if I just rest a little bit. When we read, we sit in particular positions, and that becomes part of our reading. I've done surveys with, particularly middle and high school students in Norway and in Amsterdam at international schools, and they will tell you that, when they are reading and when they're writing, that the physicality of it matters.
So the physicality of holding a real book. They'll tell you, reading is real. I mean, they're using the word, I it up. Uh, When you are holding , 20 percent of all the responses to the question of what is it that you like best about? Okay, let me get this right. What is the most important part to you of reading in a physical book? Reading in print. 20 percent of, it was 400 and some answers, , of the students were something about tactile ness. Something about the haptics, about the touch.
It's the feel of the paper. It's holding the book in my hand. It's Being able to put my fingers in and saying, I've read this much, put it this way, I've read this much and I have this much more to go. , It matters to readers that something is real. So, coming back to this question of how much we are, are consigning our thinking to an AI, it's not just the issue of the AI doing the analysis for us, and not, and are not having to learn how to do it, and to personalize it, and to keep doing it.
It's also our involvement with thinking is a bodily involvement, and that doesn't happen with digital devices, which is how, for better, for worse, we're going to access AI, I'm going to come a little bit back , if you're okay on time to talk about some of your earlier works , you've been mapping this for years, which is why.
You know, I was going to say, I know you didn't prepare to write about ChatGPT, but you'd built that muscle over years it was just a different exercise now you were talking about screens you're talking about digital you're talking about audio books and how we absorb information how we read essentially but i wanted to share you know that study of the london taxi drivers you know the famous study about their. Yeah. , so maybe I'll let you share it.
And cause I think it's, what's so interesting about this and I'll tee this up for you to maybe share with our audiences, you use it, or you lose it. That's clear from that study. And it was Robert Sapolsky who told me about this study first that it, when I first discovered it. But I often pass by. When I cycle, I'll pass by trash vans that are carrying the trash.
And I was wondering to myself, I was wondering, do those guys have an enlarged, whatever part of the brain deals with smell or turns off? I'm sure they do, you know, I was like, that would be an interesting study to do.
Cause I'm sure they're just immune to it after a while, but I share all that to say, you, talked about that, this part of the brain, and I'm going to butcher the name, they caudate nucleus, is lit up when you're going through the learning process and again, students, and we're talking about more students that have autonomy, so not.
Primary school here in Ireland primary school or secondary school students who were have oversight but the students that go through the struggle are building that neural pathway. Versus when they go to college. Then just like you said about languages, that you lose it. I used to speak fluent German. I speak fluent French still, but I lost my German because I never practiced it. And it's, that piece for me is just, it's just a shame to lose something or to never have it in the first place.
Okay, so you raised the c black car taxi drivers. This was a famous study done, and there have been other iterations of it, but I'll talk about the original one, done by Eleanor McGuire and her students, uh, in London. What they did is took people who had officially gotten the knowledge, as it's known, of going through arduous training of learning all the streets, all the sites, how to get from place A to place B. And by the way, you might want to know that the British Museum is here.
And those taxi drivers, she put into MRI scanners and found that an area of their brain, and what particularly was the hippocampus, which is responsible for navigation. I mean, it's responsible for other things, too, but particularly for navigation. And she found that the hippocampus was larger in those taxi drivers who had gone through the training to get the knowledge, and particularly had then been driving for some number.
Now we know that certain parts of the brain do enlarge depending upon what you are doing them. Um, A simple example is with musicians. It all takes string musicians and particularly violinists because I played the violin Um, And the hand that is moving the bow. is changing your brain if you go through the arduous task of practicing over and over and over and over again. You can see it in the MRI. Okay. Now, what happens with mental abilities?
You know, does Einstein have, did Einstein have a different brain? from me. And they indeed looked at Einstein's brain. There was a whole book written on Einstein's brain or whatever. It's really, really hard to pinpoint where in the brain intelligence is. So let's say you take an IQ test, if you believe that these actually measure anything worthwhile, and you've got somebody who measures an IQ of 250, which is off the charts really wonderful.
Looking at that person's brain, are you going to see any difference? You know, let's, let's assume they're not driving, you know, that this really smart person is not a black car taxi driver and is not a violinist, okay? Are you going to see any difference? There's a smidge of evidence that maybe people who, uh, read books in print versus digitally have slight differences in their brain. We don't really know this. People who are good spellers, are they smarter? Is there something in their brain?
We don't have data, not to my knowledge anyway. So it's going to be a long time in coming, I predict. Well, who knows? But, um, it's going to be a long time coming, I suspect, before we can say, well, because the lawyer used the software rather than writing the briefs himself or herself, that lawyer's brain is not as whatever. Because you have lost your German, although you kept your French, has something changed that somebody can see in your brain?
I don't know that's Um, Being bilingual by itself is great, and it changes the way that we think, we know that, but is that measurable in your brain? I don't know. So it's, it's, it's nuanced. What I do know is because I'd much rather we be living beings, moving through time and space, living our lives, thinking, writing, reading, writing, reading.
I'd like to believe that whatever the MRI scans show, we are becoming better people for ourselves and for our families and for our societies as a result. I love the idea of you were saying. When they write with the pen, people write with a pen and I often think this, I've, I've one book paled in comparison to your 11, almost 11, but I write every week and I, I feel sometimes that.
, it takes a while to get into that brainwave state that you feel like you're ready and I have to have certain conditions, et cetera. But I genuinely feel that if I actually write from that, from a different place, sometimes it, it, it's just different. It's, it's got me, it's just got something in there. And it doesn't mean that it's going to be a great article or a great book or anything like that, but it's just, it's, it's what you're trying to get out.
It's right where you're trying to get across. So I think, having gone through that, I would love people to also go through that experience and be able to do it and , not miss out on it. And it's that, that I worry about, but I wanted to change a little bit now, and this is probably.
Something that's dawned on people which is what all this ai that there has to be some impact on society that has to be some impact on jobs and this story of queen Elizabeth getting presented with the knitting machine in 1589 and , she denies the patent. And this was the days when patents was gold and this guy who actually created it was a clergyman. He went to France, eventually got the patent, but what it really highlights is.
Elizabeth whether she was right or wrong so the impact that this could have on jobs and for me that's a metaphor for regulation that you go okay, this is gonna we're not quite ready for this yet and i'm gonna link it then to something that you said that the US department of labor estimates that american writers, add copyright writers grant writers speech writers and more collectively earn, i could not believe this figure six hundred and seventy five billion
dollars every year that's a huge amount of taxes paid to the government but it's a livelihoods for so many people and that doesn't include people who write in there as well the gig work that you can't account for so easily and. This speaks to the idea that you talked about paralegals if i'm working in it that those jobs are just being wiped out and i'm less of a tech optimist than i used to be because i used to see that in the past.
Each technological jumped of each revolution created more jobs now i hope one thing i hope that i am like the person not like a luddite but i'm the person who is just not seeing how this is gonna create more jobs but i just can't see it because the businesses that are dominating today are platform businesses.
And platform businesses are measured on revenue per employee and if that's high, so they've less employees per higher revenue their valued more wall street happier so the only way people gonna make money from that is if they have got stocks in those businesses. So it's, it's a, it's a real concern, but also a dilemma. And I wondered what you thought about that.
Alright, the first thing I will say is you talked about how much revenue according to the U. S. Department of Labor is generated by people who have the classification Writer. Look at the year on the data. It is not 2024. Those data will take a couple years to run, so I will change from earn to earned. Okay, so this is making me sound as if I'm pessimistic. Am I pessimistic?
Well, I will take as an example someone who just won a Nobel Prize in Chemistry by the name of Demis Hassabis, who was co founder of DeepMind, which was first known for creating, with a game called AlphaGo. which bested a Go master, the ancient game of Go. And then he created something called AlphaFold. And it was for AlphaFold that he got the Nobel Prize. Now it wasn't that specific item, but AlphaFold is a program that was created in order to determine the structure of proteins.
And what he said was that after AlphaFold was shown to be so good, is, this is why I did the games. Not because I was interested in Go and being able to best a human being. That was fun. , and Hasabas knows a lot about gaming. He was producing video games from way on early.
, But he wanted to do something that would change the capabilities we have in the field of medicine and health in order to, two, better understand the structure of proteins because deciphering, unfolding proteins, the old fashioned human way, took years to do a single protein. But now if you can do it with AI, there are all these new kinds of drugs you can invent, there are all these new kinds of diseases you can figure out a way to attack.
in the sense of So it seems to me, one way to think about the promise of AI is not to ask, how can it do the things that humans can do, so you don't know if a human did it or whether an AI did it. but to ask, how do we really turn this on our head? Which we started talking about after ChatGPT came online. Namely, how do we augment what humans do?
Well, the problem for the augment discussion for the last two years is every time we said, but now only the human can do such and such, well, the AI caught up. So we don't know right now what it is that humans should be looking to do themselves. But it seems to me that the most important thing we can do in education, the most important thing we can do in industry, is to ask, what is it that only humans can do that will also benefit them? So if AI catches up, fine.
What else is it that only humans can come up with? What kinds of ideas? We know that AI can be really useful for brainstorming. And there are lots of people who have said from the beginning when ChatGPT came online, you know, use it for brainstorming. That's great. Don't have it do your work We need to add to this an ethical component that people have to care to create for themselves and to know for themselves and to express who they are.
In their writing, or in their other forms of creativity, it has to matter to them. When it comes to writing, and I've written about this a lot in the last two years, and if higher education, and I'm speaking for the United States more than anything else, if higher education doesn't care to devote the resources of the faculty, to work individually with students, to talk about their ideas of what they might want to write, to look at a draft, to go through it together, to then hone it together.
And that's very labor intensive and it costs. Right? And most universities and colleges in the United States have one or two English comp classes and then it's over. And then the student's quality of writing traditionally has declined after that because the teachers don't care about the writing, they're not grading it anyway. If we don't care to nurture writing, why should the students care? Okay, so there's, there's a larger puzzle here.
It's not all on the learner to motivate himself or herself. There has to be an infrastructure. That cares as well. I think that's a really nice way to conclude today. What do you think? I'm done. I mean, this is this has been a lot of fun, but I think I've given you enough food for thought And then I would never say that Naomi and AI would keep talking. and I would not be asked its opinion in the way that there was a mind that could decide.
was absolutely brilliant let's share where people can find you . Your books where can people find you. My American University webpage is probably the simplest way. I'll link to that. Naomi as well. And when will the new book be out? My hope is that a year , it'll at least be done. What happened with the older book, and I think I told the story in the book, I don't know. I had submitted the final manuscript for who wrote this in mid November 2022. Two weeks later, Chat GPT came out.
And I am incredibly grateful to my editor and to Stanford University Press because the plan had been for my book to come out in late, in, in spring 2024 for who wrote this. And they said, you know what? It needs to come out now. And , I rushed the revisions, and they rushed the production, and they got it out in September 2023. And, , how that, , dance I can do with the new book? We'll see.
i'm gonna link to, where to find you again i hope to have you back on when you have the new book it's been an absolute pleasure author of who wrote this, Naomi S. Barron. Thank you for joining us. Thank you for having me. It's been fun talking.