have long been among the characteristics of humanity as a species.

But today's guest chronicles the rise of artificial intelligence and its myriad abilities to write, to compose, to create, and what it means for our humanity.

She's Naomi S. Baron this week on "Story in the Public Square."

(bright music) (bright music continues) Hello and welcome to "Story in the Public Square" where storytelling meets public affairs.

I'm Jim Ludes from the Pell Center at Salve Regina University.

- And I'm G. Wayne Miller, also with Salves Pell Center.

- And our guest this week is Dr. Naomi S. Baron, professor Emerita of Linguistics at American University.

She's also the author of an important new book "Who Wrote This?

How AI and the Lure of Human Efficiency Threaten Human Writing."

She joins us today from Washington DC.

Naomi, thank you so much for being with us.

- A pleasure to be here.

- You know, so for audience members who might not know what the field of linguistics is, would you explain that for us?

- In 50 words or fewer?

(Jim laughs) Very simply, linguistics is the study of language.

Generally that means human language, but an awful lot of my colleagues have worked on animal communication as well.

So what constitutes language?

Well, there's spoken language, there's written language, there's sign language, and these days there's also language that is generated by artificial intelligence and interpreted by artificial intelligence.

So that largely covers the field, I think.

- And is that how it's, through the prism of linguistics, that you came to study artificial intelligence in the first place.

- Right, and to be fair, people who have been interested in artificial intelligence from the very coining of the phrase, artificial intelligence, essentially in 1956, primarily we're interested in how language can be used as a measure of whether a computer is capable of doing what a human could do, that is capable of human intelligence.

Early on in the career of artificial intelligence, especially in the United States, there was a lot of work on machine translation from one language to another, essentially Russian into English.

This was called war days.

So that's language, again.

So this is issue of language.

So-called natural language processing, which has to do with generating and understanding human language, has been at the core of an awful lot of AI initiatives from the get go.

- So, obviously your book is about writing, and we're gonna get into the book in depth momentarily.

But here's a very general question, and actually one I'm really looking forward to the answer to as a writer myself, why do people write?

- There are many, many, many reasons, and there are individual reasons.

There are individual reasons modular of profession.

So you may write because you want to earn a living, and that's one way to do it.

You may write because you just want to keep track of things that you are thinking.

Think of a diary, think of a blog post.

Writing actually evolved a number of different times historically.

Some of the earliest developments of writing in what is now Greece, this was a long time even before Plato and Aristotle, was for keeping records of what the palace administration, this was Mycenaean times, think of King Agamemnon, you know, and the Iliad.

Writing was essentially evolved.

It was called linear A, linear B, rather, for keeping track of how many grains of wheat were going into the castle storehouse, and so forth.

Other kinds of motivations are to share your thoughts with other people.

So people in literature often do that.

Sometimes your motivation is to vent, because you're really angry about something.

Think about some letters to the editors or to friends or to employers whose employment you're about to quit.

So the range of reasons people write is huge.

And to segue into why I care about motivations for writing, it's that if AI does a lot of the writing for us, I worry that we reduce our own human motivations for writing, because those motivations for writing are also motivations for thinking, for understanding ourselves, and for forming social communities.

- Does writing alter our brains in any way?

- Yes.

We know that writing evolved actually a number of different times, beginning, we think, around 6,000 years ago.

But evolutionarily, that's not nearly enough time for writing to be wired into our brains.

So what happens when we learn to read and to write is we rejigger the kinds of neurological networks largely that were developed for spoken language, because those are now evolutionarily wired into us.

We rejigger them so that we can deal with a new form of linguistic representation.

There's some really interesting experiments that have been done comparing adults who are not literate with those who are, you do MRI scans of their brains, and then you teach them how to read and write and you do new scans and you see their brains have changed as a way of accommodating to needing to cope with reading and writing.

- I guess I wanna ask this sort of the same question, but maybe from a slightly broader perspective.

- [Naomi] Okay.

- Is writing one of the things that makes us human?

- That would not be a fair thing to say for a couple of reasons.

Number one, there are lots of people who are non literate.

So if you take a 4-year-old, is that 4-year-old not human?

- [Jim] That's a great point.

- You're not a human until you're an adult.

But we could get on with the story.

There are many peoples who are not literate.

They're surely human societies before, you know, let's say 7,000 BC, 7,000 years ago, rather.

People were human.

The question is, what does literacy do for us?

And one of the things it does for us is it gives us a way of literally seeing what we've been thinking about, being able to reevaluate what our thoughts are, to have other people look at them and critique them.

The arguments have been made that, with the development of the Greek alphabet around, oh, probably around 700 BC, that it became possible over the next several hundred years for Greek philosophy as we know it, classical Greek philosophy to develop, because so much was now being written, looked at, critiqued, redone, and those are skills that add to our humanity.

But I would be the last person to say that if you're not literate, you're somehow not human.

- Yeah, that's a great point.

So, you know, we wanna talk about artificial intelligence and its impact on human writing and writing more generally.

Maybe we should start though with a definition or an explanation of what artificial intelligence is, and specifically the kind of artificial intelligence that we're talking about when we're talking about writing by AI today.

Could you walk us through that?

- Sure.

There would be many definitions of artificial intelligence, but the simplest is use of some computer-based technology to accomplish what a human intellectually, cognitively could accomplish.

So you're looking at the results, you're not looking at the method via which the AI, or the computer programming, the algorithm, and then the algorithms, and then the software, and the computer chips that are driving all that computation happening.

You're not looking at the process via which it's happening so much as, have you gotten a result that is indistinguishable from what a human could do?

There's the famous, so-called Turing test that Alan Turing devised.

He called it the Imitation Game, this was back in 1950 when he said, "If you could get a machine to produce language," and all of this was written at the time, "such that a human being couldn't tell if that language was produced by a human or by a computer, that would be," not his term, because the term artificial intelligence hadn't been coined yet for another six years, but "that would be evidence of a machine being intelligent."

Again, you're looking at the results and can you distinguish those results from what a human could do.

- So back in those early days, why were the people who were working on machine writing, why were they interested in having a machine do what people could already do?

- All right, let's go back to Alan Turing.

During World War II, he became known for, along with his team at Bletchley Park in what we now call the UK, we used to call England, he became known for decoding German submarine codes to figuring out, you know, the so-called Enigma codes, which are actually many.

And although his training was really as a mathematician, he became a code breaker.

Language is also a form of code.

And some of the early discussions in artificial intelligence on both sides of the Atlantic talked about using computers, which indeed Turing and his colleagues had used for breaking the Enigma codes, for using computers to break another kind of a code, which is, how is language made up?

How could you replicate what human languages can do?

So a lot of the early people doing work in AI were mathematicians, and mathematicians, as we knew from World War II, again, on both sides of the Atlantic, made really great code breakers.

So the two sort of natural code breaking and language, and then you throw in using computers to do that code breaking really are natural partners.

- So when we are, you know, I think a year ago I was grading my last set of papers for the semester and a friend of mine sent me a link to something called ChatGPT.

And I spent, I don't know, an hour or two that first night sort of going down that rabbit hole and being both amazed and a little bit horrified at how effective OpenAI's platform was at producing human sounding language.

How big of a revolution is this and what does it portend for what's still to come?

- Okay, two questions.

How big of a revolution, the big revolution was it millions of people suddenly became aware that it was possible to have computer algorithms generate written language that looked pretty much, not fully, but pretty much indistinguishable from what a human could produce.

This kind of technology was already out there.

It was out there even before OpenAI started producing GPT.

It didn't label it one, but it was one, then GPT-2, then GPT-3, then ChatGPT is GPT-3.5.

We now have GPT-4 from OpenAI.

So the technology existed before ChatGPT was launched for free by OpenAI to the world.

OpenAI was really interested in getting a lot of free user data, which is why they made it freely available.

They didn't have any clue so many millions of people would be interested.

Okay, but millions of people were interested and it seemed a nifty tool both for writing essays and for writing analyses of whatever.

And then other kinds of tools have been added to what ChatGPT can do, especially if you pay your $20 a month to OpenAI for using more sophisticated features.

What does this portend?

It's really impossible to know yet.

Remember, ChatGPT was released on November 30th, 2022.

That's a second ago in the time of evolution for how we use computer tools of any sort, including AI tools.

What's pretty clear in the world of education is that there was first panic, "OMG, what's gonna happen?

I won't have any clue if my students have written the papers."

Then there was development of a lot of tools, probably GPT-0 by Edward Chung and his colleagues now is the best known of the tools for trying to detect if something was written by either a ChatGPT or one of the other GPTs, the Generative Pre-Trained Transformers as the initial stand for.

And then educators started saying, "You know what, if you can't lick 'em, join 'em.

So let's see what productive uses we could make of this technology."

So for example, have ChatGPT or some other GPT generate an essay and then have students evaluate it or have students write an essay and then have a GPT evaluate it.

So there are various kinds of, some would call them innovation, some would call them games, that a lot of people in faculties and universities and colleges as well as in high schools are playing with right now.

We do not know where this is going to go.

What is clear is that an awful lot of students are saying, and some faculty members as well, the most important change that's gonna come about as a result of this technology is we've gotta rethink educationally, pedagogically what kind of assignments we want to make to students.

Do we just wanna play the game of gotcha or is there something else we could do when we're asking students to write?

- So let's have you get into the implications for some of the writing professions starting with starting with journalism.

- I was hoping you'd do that.

(all laughing) - I'm a journalist, so I really need to know the answer here.

- Okay.

Starting with journalism.

Journalism was actually, or some form of journalism, namely writing quarterly reports for Fortune 500 companies, was one of the very first uses of AI to generate articles.

And the reason that was so easy to do is those quarterly earnings reports are total templates.

I mean, you say the same thing about every company, you just fill in the numbers differently.

The same thing was true of giving reports on basketball games and baseball games, which is again, where some of the technologies started.

The story with the journalism as I see it now, is twofold.

Threefold, sorry.

Let's go for threefold.

The first is an awful lot of professors of journalism are saying, "We need to be gung-ho about the digital technologies for journalism because you know what?

We're now gonna free up journalists for writing the feature stories, for doing investigative journalism, and it's really gonna be wonderful.

We're not gonna see any diminution in the number of jobs available."

That's theme one.

Theme two is that maybe yes, maybe no, it's too soon to know.

There are indeed lots of newspapers and magazines that are retrenching, saying "AI can do this and they can do it more effectively in terms of our finances."

But number three, the newspaper industry, I'm sad to say, has been undergoing big time retrenchment long before generative AI came along.

As a lot of people, particularly with the rise of digital technologies and reading newspapers online and having things tweeted and retweeted, or X-ed or re-X-ed, depending on how you wanna describe it, are getting their news from social media or getting headlines from a newspaper that they can access for free.

And as a result, it's both the big national newspapers as well as local newspapers especially that are retrenching or closing entirely.

So it's really hard to know how much of the blame for the retrenchment of journalism, it's not the death of journalism, the retrenchment of journalism, what's happening anyway, and how much will be a result of AI being able to generate really good stories, increasingly good stories, increasingly analytical stories, not just plain vanilla fluff, in the years to come, the near use to come, I might say.

- Yeah.

So, Naomi, the concept of an artificial intelligence platform, crafting stories based on an established set of facts or understandings is one thing in my mind.

But what about creative?

What threat, what challenge does AI pose to more creative pursuits?

Whether we're talking about a work of fiction or the composition of a symphony, for example.

- Okay, so let's work on fiction first.

There are a number of writers who have said, "You know what?

I know that AI could do good stuff, but it's not there yet.

So I'm not worried."

One of the most recent worries, and I've been talking with actually a couple of authors, very well-known authors, bestselling authors, especially in Europe, who are having their text fed into AI data sets and things written in their style.

So if you wanna talk about a paragraph or a page worth of writing, that sounds really good and sounds creative linguistically, that's already doable.

And by the way, a number of these authors, including Margaret Atwood, are suing open AI because this kind of thing is happening, and other companies that have large language models as well.

But the larger question is, well, what about the plot structure?

You know, we thought, you take a really intricate plot, you take "Anna Karenina," could degenerative AI come up with the plot, the plots, I should say of Tolstoy and of his novels?

And the answer at this point is probably no, but we don't know where things are going to go.

To me, the most much more important question is not will it be possible next year, five years from now to create a really intricate piece of literature that we cannot discern whether it is AI generated or human generated, but it's something people would like to read and looks really creative.

That to me is less the question than what is human creativity about.

Because an awful lot of what matters to us as human beings is that something was created by a human being.

Take an analogy.

If you really want the best recording possible of, oh, what should we take Mozart's 29 Symphony.

You can do that in a studio with the Berlin Philharmonic and have a magnificent recording.

If you go to a concert that the Berlin Philharmonic is playing with the same orchestra, maybe they mess up just a little bit.

And if you're really good violinist, you notice it.

But what matters is that you're seeing human beings, live human beings making music.

So I think in the years to come, the discourse is going to change from, oh, could AI be creative the same way in terms of the same results that a human being could have.

I hope it will change, I should say, to, we care what human beings can do.

And if an AI can do something that looks equally good, more power to the programmers, but I still want humans to do the creating, 'cause it changes who we are when we create.

- So I think that brings us quite naturally to two strikes this year involving Hollywood: Writers and actors.

What are the implications for Hollywood, and obviously the actor strike is continuing as of this taping?

What are the implications for Hollywood?

I mean, that was a huge point and remains a huge point in negotiations.

- Well, I read the largest number of details I could find on what the AI clauses are for the screenwriters.

And the takeaway is, if AI is gonna be used, you have to let us know, and we have to let you know.

That is, you have to be transparent about where things are being used.

It doesn't say thou shalt not ever use AI for anything, because quite frankly, that's unrealistic.

Where is this going to go?

The discussion among people in the AI community who worry about these things goes like this.

We need to augment human abilities, not automate them.

There's a business professor at the Wharton School at the University of Pennsylvania who describes what should happen as creating a writing centaur, that is part man, part beast, in this case, part computer, part human, and that we need to learn to harness the things that generative AI can do for writing, but make sure the human element is never lost.

That it's more a collaborative partnership.

It's an augmentation rather than a replacement.

That said, I'm not convinced that the number of jobs is going to remain whatever level it is now, because augmentation saves lots of time and time is money, not just in Hollywood, but in other places as well.

So I think the contract was probably the best one one could get.

What happens the next time a contract is negotiated as AI has progressed enormously in the next several years, which it will, I don't know, but I'm a little more pessimistic than people were a year ago about saving all the jobs that we say we can save.

We say, oh, we're gonna have people who are good at being prompt engineers, and isn't that a wonderful profession?

Well, not everybody can be a prompt engineer and not everybody's running the software.

So people I would like to hope still have the creative spark within them to write, but I'm hoping the market and the contracts make it possible for humans to keep doing so.

- Hey, Naomi, we've got about a minute left and we could talk to you for another week, I think, but you know, as I've processed this myself, I sometimes wonder, am I just being a Luddite?

Is my resistance to change to the adoption of new technology, what I see my students doing and so forth?

Is that just simply a romantic notion of things used to be better and why can't we just keep things well enough alone?

Where do you come down on that?

- Okay, and I'll give an analogy, and that is reading books in print versus reading them digitally.

Digitally is technology, right?

The most recent report from the American Library Association shows that Generation Z readers, students are more likely to read in print than to read digitally, and generation, you know, the Millennials are still more likely to read, are more likely than Generation Z to read electronically.

So the younger users, you'd say younger, they've gotta be more digital, right?

They've gotta want more technology.

No, that does not seem to be what's happening, at least with reading.

We're gonna see what happens with AI.

I think a year from now, if we had this conversation, things will have calmed down.

We'll find some modus vivendi, and I hope there'll be enough people arguing for the importance of humans continuing to write, which means also to think and create.

- Naomi Baron, the book is "Who Wrote This?"

It's an important read.

Thank you so much for being with us.

That is all the time we have this week.

But if you wanna know more about "Story in the Public Square," you can find us on social media or visit pellcenter.org where you can always catch up on previous episodes.

- For G. Wayne Miller, I'm Jim Ludes asking you to join us again next time for more "Story in the Public Square."

(bright music) (bright music continues) (upbeat music) (no audio)