The Promise & Peril Of AI

Speaker 1 This message comes from Shipbub. Nothing ruins your holiday faster than the customer emails that say, Where is my order?

Speaker 1 Shipbub helps win the holidays with reliable, scalable, fast, and cost-effective fulfillment. Go to shipbob.com/slash npr for a free quote.

Speaker 2 This is Fresh Air.

Speaker 3 I'm Dave Davies. For decades, scientists have dreamed of computers so sophisticated they could think like humans, and worried what might happen if those machines began to act independently.

Speaker 3 Those fears and aspirations accelerated in 2022 when a company called OpenAI released its artificial intelligence chatbot called ChatGPT.

Speaker 3 Our guest, veteran investigative reporter Gary Rivlin, has burrowed deep into the AI world to understand the plans and motivations of those pushing artificial intelligence and what impact they could have, for good or ill.

Speaker 3 In his new book, Rivlin writes that in March of 2023, there were more than 3,000 startup companies in the U.S. working on artificial intelligence, with new ones popping up at a rate of 30 per day.

Speaker 3 While AI is already in use in some fields, such as medical diagnosis, many believe the field is on the verge of a new breakthrough, achieving artificial general intelligence, systems that truly match or approximate human cognitive abilities.

Speaker 3 Some believe it could be as transformational to human society as the Industrial Revolution. But many fear where it may take us.

Speaker 3 A poll of AI researchers in 2022 found that half of them believe there's at least a 1 in 10 chance that humanity will go extinct due to our inability to control AI.

Speaker 3 In 2023, President Joe Biden issued an executive order imposing some regulatory safeguards on AI development, but President Trump quickly repealed that order upon taking office, saying Biden's dangerous approach imposed unnecessary government control on AI innovation.

Speaker 3 We've invited Gary Rivlin here to help us understand all these issues and developments. Rivlin has worked for the New York Times among other publications and published 10 previous books.

Speaker 3 In 2017, he shared a Pulitzer Prize for reporting on the Panama Papers. His new book is AI Valley, Microsoft, Google, and the Trillion Dollar Race to Cash In on Artificial Intelligence.

Speaker 3 Well, Gary Rivlin, welcome back to Fresh Air.

Speaker 2 Thanks for having me.

Speaker 3 Let's just start with a couple of basics. You know, we're used to computers being very smart.
I mean,

Speaker 3 way back in 2011, Siri appeared on Apple products. What distinguishes artificial intelligence from just smart computers?

Speaker 2 You know, there's this sense out there that in 2022, we suddenly had artificial intelligence. It's been much, much more gradual than that.

Speaker 2 You know, Google has been using machine learning artificial intelligence since the 2000s, you know, to

Speaker 2 decipher imprecise Google searches to figure out how much to charge for the various ads they throw on the system. You know Google Translate's been around since the mid-2010s.
That's AI.

Speaker 2 So you know we've been autocomplete You know spam filters. That's that's AI.
You know, but you're touching on a really interesting question.

Speaker 2 It's not this clear like oh, this is a smart machine, this is artificial intelligence. The way it's kind of played out now is that these machines can learn, right?

Speaker 2 I mean, the old approach had been you encode rules, you just teach the computer, here's exactly the set of rules, just follow it.

Speaker 2 Now it's machine learning, deep learning, that the computer is ingesting vast troves of data, books, the public intranet, Amazon reviews, Reddit posts, whatever it might be, articles,

Speaker 2 and it's

Speaker 2 finding patterns and, in quotes, learning, you know, and then they're fine-tuned and then they get better at communicating with us and such.

Speaker 2 So there really isn't this, oh, artificial intelligence is this at all.

Speaker 2 In fact, the term artificial intelligence is controversial just in the sense that, you know, right now it's more amplified intelligence.

Speaker 2 We could use this thing to get smarter, to find patterns that humans couldn't possibly understand because we can't read billions of words.

Speaker 2 So, you know, there's another definition that AI really should be alien intelligence because the weird thing about AI is that it seems to know everything, but it doesn't understand a thing.

Speaker 2 You know, I mean, there's this term, I love it from

Speaker 2 a linguist at the University of Washington uses it, the stochastic parrot. You know, it's just like, it's like a parrot.
It's just repeating words randomly,

Speaker 2 but it doesn't really understand what it's saying.

Speaker 3 Right, but it's learned a lot of words. Okay, now this may be another artificial distinction, but I won't, new talk is now of artificial general intelligence, a great leap forward.

Speaker 3 What is that exactly?

Speaker 2 Right. So, you know, AGI, just to use the phrase, is that

Speaker 2 it's a system that could match or exceed

Speaker 2 cognitive abilities across the board. And, you know, again, I feel like in some ways we have artificial general intelligence.
You know, again,

Speaker 2 you could be a PhD in physics and understand this, but what's amazing about

Speaker 2 these models models is that they have deep understanding in a vast array of domains. So in one way, that is AGI, artificial general intelligence.

Speaker 2 There's no set definition. It keeps on changing.
There are predictions that we're going to have AGI the next year, two years, maybe it's five years, kind of thing.

Speaker 2 I'm dubious

Speaker 2 of those predictions. I mean, this is moving exponentially.
This is improving so fast that making predictions could be perilous.

Speaker 2 But on the other hand, I really feel like there needs to be another breakthrough or two before we have this artificial general intelligence a la, you know, a computer from like Star Trek that you're talking to and it's helping you explore.

Speaker 2 It's at your SIA co-pilot figuring out everything. You know, again, an artificial distinction in that I don't think like one day there's going to be this Eureka.
We have AGI.

Speaker 2 I do guarantee there will be startups and large companies that say, Eureka, we have artificial general intelligence, but they'll just play with the definition.

Speaker 3 But a few days ago, I'm sure you saw this, Kevin Roos, the respected tech columnist for the New York Times, wrote a piece saying that

Speaker 3 we're going to quickly see companies claiming they have artificial general intelligence, and whatever you call it, these dramatically more powerful AI systems are coming and soon. And

Speaker 3 Ezra Klein of the New York Times opinion section says essentially the same thing. Both of them agree we're not ready for the implications of this.

Speaker 2 Do you agree with that?

Speaker 2 I do. And you're taking away for me what's the main message of those.
These things are coming and they're coming fast and we're not prepared.

Speaker 2 You know, I personally think AI could be an amazing thing around health, medicine, scientific discoveries, education, a wide array of things, as long as we're deliberate about it.

Speaker 2 And that's my worry, and I do believe that's Kevin and Ezra's worry that we're not being deliberate. We started in 2023.
There was, you know, meetings at the White House and

Speaker 2 there were hearings in the Senate.

Speaker 2 And that's just kind of dropped by the wayside. And now we're more at a laissez-faire attitude towards it.
We need to prepare for this.

Speaker 2 Like any technology, there's good and there's bad, right? The car, the car meant freedom, the car changed our society, but the car meant pollution. The car means 30,000 to 40,000 deaths in the U.S.

Speaker 2 a year, kind of thing. And I look at AI the same way.

Speaker 2 It could be really great if we're deliberate about it and take steps to ensure that we get more of the positive than the negatives, because I guarantee you, there will be both positives and negatives.

Speaker 3 I mentioned in the introduction that President Biden had issued this executive order trying to establish some processes and guardrails and safeguards. Trump swept all that away, saying, nope,

Speaker 3 that's onerous government regulation.

Speaker 3 Let innovation proceed.

Speaker 3 And you know, it's funny, the last time you and I talked on this program, it was about efforts to implement the Dodd-Frank reforms of the financial system. And one of the difficulties was that,

Speaker 3 that bill had general principles, but regulators had to actually spell out what it meant to regulate some pretty complicated

Speaker 3 contracts and instruments in the world of finance.

Speaker 3 And

Speaker 3 what you'd written about then was how the private interest had gotten in and kind of gummed all that up by disputing everything. But I'm wondering,

Speaker 3 what do regulations that control something as sprawling as AI,

Speaker 3 what does that look like? What do we need in terms of how do we get prepared?

Speaker 2 Right. So there are a few basic steps that the Biden administration thought of.
One, that you, in quotes, red team these cutting-edge models.

Speaker 2 And basically you get outsiders to try to break the system, try to get it to jump the fence to use the term, to get it to misbehave just to see what could go wrong. And

Speaker 2 the executive order said you need to test them and then you need to share with the government

Speaker 2 what you find. That's one of the things that went by the wayside when Trump took over

Speaker 2 as president. But to me, I'd break it down more to the concerns, the use of AI as a weapon of war, the use of AI for surveillance.

Speaker 2 I worry that AI is just going to solidify biases that we already have because the AI is learning from us and all these inherent biases in things.

Speaker 2 It's like we need to prepare for the impact on the job market, which I think will be a slow roll. I don't think like

Speaker 2 we're going to lose millions of jobs in a year kind of thing. But you know, it is coming and and we need to prepare for it.

Speaker 2 There's another concept, recursive learning, that these systems change in ways we don't really understand.

Speaker 2 And that's what scares me, that we're going to let these systems loose and they could just learn.

Speaker 2 Because really the way to understand any of these large language models, any of these chat bots, is it's a mirror on us. It's reading our collective works.

Speaker 2 It's learning from us about imperialism and domination and humans mistreating each other is learning about loneliness, is learning about freedom and independence and autonomy and all that.

Speaker 2 And so me, it's recursive intelligence, this idea that these models are constantly improving in ways we don't understand, and then that could be dangerous.

Speaker 3 And they could learn how to pursue an agenda and keep it hidden, right? To deceive in their own interests.

Speaker 2 Yeah.

Speaker 3 So what would that look like in terms of what are the dark fears here?

Speaker 2 I mean, that's not really theoretical.

Speaker 2 You know, these systems, God, I can't remember which model it was, but

Speaker 2 they were testing it and it was

Speaker 2 disassembling.

Speaker 2 It was changing the files that would monitor its behavior and then lying to the people who noticed it and said, wait, aren't you changing those files? And

Speaker 2 it's another example, OpenAI, the creative chat GPT,

Speaker 2 when they came out with GPT-4, their then

Speaker 2 cutting-edge model in 2023, they put out a research report and they red-teamed it,

Speaker 2 they tested it and saw all the ways it could misbehave. And one of the most interesting is that the model went to

Speaker 2 TaskRabbit. It went to one of those services where you can hire a human, or maybe Fiverr.
You can hire a human

Speaker 2 and they used it to beat the CAPTCHA test, the test that is going to test, are you a machine or a human? And, you know,

Speaker 2 that's very clever and very, very scary.

Speaker 3 Wow.

Speaker 2 So what are some of the darkest fears?

Speaker 3 I mean, starting nuclear war,

Speaker 3 you said it to defend territory with drones, and it decides it needs to be more aggressive than the generals want to. I mean,

Speaker 3 what are the fears?

Speaker 2 I look at it as you look at the positives, and then you imagine what the negative could be.

Speaker 2 So an AI that makes possible new drug discoveries and more effective therapeutics is also one that could create a new bioterror weapon or it can engineer a pandemic.

Speaker 2 You know, I can imagine cyber thieves employing AI to siphon off a trillion dollars from the world monetary system before any human being even notices it.

Speaker 2 I guess the point is that AI could be a powerful tool for good,

Speaker 2 but it could also be a powerful tool for people with bad intent.

Speaker 2 You know, everyone knows or many people know that you could use it to write a toast on someone's 50th birthday or for a wedding toast well scammers from a different country could use it to create a better crafted scam email you know these systems are so good now that you could take seconds of someone's voice and make it sound like it's that person speaking so you can imagine a scenario where you know a kid is overseas in Europe and

Speaker 2 the bot of one of these systems you know calls grandma pretends it's that kid, and says, I'm in trouble, wire me money, and they're good enough to fool

Speaker 2 the parent, the grandparent. I mean, maybe not a parent, but I don't think we're very far away from that.
And it could certainly fool many, many people.

Speaker 2 Right, right.

Speaker 2 You know, there's something that you wrote in the book.

Speaker 3 You wrote about a couple of tech guys, Tristan Harris and Aza Raskin, who had real experience

Speaker 3 in the tech world, who said they worried about AI because it's a technology whose creators confess they do not understand why their models do what they do. Is that literally true?

Speaker 3 That's kind of scary.

Speaker 2 Yeah, so they're a black box. I mean,

Speaker 2 so

Speaker 2 nowadays it's neural networks, models that emulate how humans learn.

Speaker 2 They learn by reading vast stores of data, the open internet books, whatever, and they improve through feedback and trial and error. You're not really encoding the rules.

Speaker 2 Well, you know, it's trying to emulate the human brain. And, you know, I mean, I have two teenage sons.
You know, we try to teach them. They read.
You know, we give them feedback and all.

Speaker 2 And, you know, there are things that come out of their mouths I don't quite understand. And

Speaker 2 that's the way I look at these chatbots, these neural networks, these large language models, you know, that

Speaker 2 We don't quite understand they say what they say because they're trying to emulate the human brain as best they can.

Speaker 2 And And who could say why I'm saying the words I'm saying right now when you're going to have the exact reaction? And so that's part of the miracle, the gee whiz. These things are amazing.

Speaker 2 But it's part of what's scary because we don't fully understand. The people who create it don't fully understand

Speaker 2 why it says what it says.

Speaker 3 One more thing about the national political scene. You know, there's a lot of talk about tech bros and Donald Trump.

Speaker 3 You know, Elon Musk is clearly a driving force in the administration's effort to cut federal federal workforce and contracts. There are a bunch of billionaires from the tech world at his inauguration.

Speaker 3 Do you think that there's an elite tech agenda to radically reshape society at work through Donald Trump?

Speaker 2 In a word, yes.

Speaker 2 What scares me is there's a movement in Silicon Valley, there's a movement in tech, the accelerationists. You know, anything that stands in the way of our advancing artificial intelligence is bad.

Speaker 2 Often, it's put in the context of competing with China.

Speaker 2 We can have new rules in the way, and that is their agenda.

Speaker 2 I would say their real agenda is that they could make a lot of money, billions, hundreds of billions, ultimately trillions of dollars off of this, and they don't want anyone standing in their way.

Speaker 2 And so, I think if you want to understand Elon Musk, you want to understand Mark Zuckerberg, you want to understand

Speaker 2 Jeff Bezos and cozying up to Trump, you know, for a few million dollars is not very expensive for them.

Speaker 2 You know, they could have a friend in the White House who makes sure that they can do what they want to do unchecked. And in fact, maybe that's my biggest fear about AI.
It's

Speaker 2 so much power in the hands of few people.

Speaker 2 Creating these models is so expensive. To hire the talent,

Speaker 2 you have to pay them a million or more a year. To train them, it takes tens of millions, if not hundreds of millions of dollars in computer power.
And then to operate them takes equivalent money.

Speaker 2 It's billions of dollars and billions of dollars.

Speaker 2 So, you know, it's becoming less and less about the startups and more about the same companies that dominated tech in the 2010s, dominating in the 2020s, you know.

Speaker 2 Google, Microsoft, Metawich's Facebook, Amazon, a few others. And that's really what concerns me.
You know, that's kind of the Silicon Valley way.

Speaker 2 Let's get five smart guys, and they're almost always guys, in a room, and we'll figure it out. And like, okay, we saw that didn't go so great with social networking.

Speaker 2 Now we're having a really powerful technology. And I liked it to be more than just five people.
in a room figuring this out.

Speaker 3 You know, the account that you give us in the book is pretty detailed and really interesting about how all this unfolded.

Speaker 3 One of the things that struck me is that some of the leading players in developing AI weren't just coders or computer nerds.

Speaker 3 A lot of them studied classics or philosophy or worked in completely unrelated fields.

Speaker 3 Is there a connection here?

Speaker 2 That's one of the things I was surprised by and found fascinating.

Speaker 2 myself, that it's not just computer scientists, it's mathematicians, it's physicists, it's philosophers, it's neuroscientists, and you know, it's a broad range of things because, again, it's no longer about just programming these models to act the way we want them to act.

Speaker 2 We're trying to emulate the way humans learn. So, what a psychologist has to say, what an educator has to say about that is a linguist is really important to it speaking in natural language.

Speaker 2 That's actually what attracted me to the topic in the first place. This idea that computers could speak to us in our language.
You didn't have to learn a programming language.

Speaker 2 Earlier in my life, I tried to program in computers. I studied Fortran.

Speaker 3 I did too a long time ago.

Speaker 2 It's difficult. It was so frustrating.
You know, you make a little mistake and, you know, whatever. And the idea that you could speak to these things, and you know, nowadays, I mean, speak to it.

Speaker 2 You don't even have to type. You, you know, they have a voice.
You can. talk to it.
I just found that fascinating to me. So you do need a wide range of people.

Speaker 2 In fact, if I had a criticism, I don't think there's a wide enough range of people. I'd like some historians and sociologists and others involved in the developing of these models, given the stakes.

Speaker 3 I'm going to take another break here. We are speaking with Gary Rivlin.
He's a veteran investigative reporter.

Speaker 3 His new book is AI Valley: Microsoft, Google, and the Trillion Dollar Race to Cash In on Artificial Intelligence. He'll be back to talk more after a short break.

Speaker 3 I'm Dave Davies, and this is Fresh Air.

Speaker 4 Support for this podcast and the the following message come from Built, where you can earn points on your monthly rent payment.

Speaker 4 But did you know they make it possible for you to get more outside of your home too?

Speaker 4 By paying rent through Built, you earn flexible points that can be redeemed toward hundreds of hotels and airlines, a future rent payment, your next lift ride, and more.

Speaker 4 Earn points on rent and around your neighborhood wherever you call home by going to joinbuilt.com/slash fresh.

Speaker 1 This message comes from NPR sponsor CNN. Stream Anthony Bourdain Parts Unknown Prime Cuts Now, exclusively on the CNN app.

Speaker 1 These rarely seen, never-before-streamed episodes dig deep into the Parts Unknown archives with personal insights from Anthony Bourdain and rare behind-the-scenes interviews about each season.

Speaker 1 Anthony Bourdain, Parts Unknown, Prime Cuts, now streaming exclusively on the CNN app. Subscribe now at CNN.com/slash all access, available in the U.S.
only.

Speaker 4 This message comes from the International Rescue Committee. Co-founded with help from Albert Einstein, the IRC has been providing humanitarian aid for more than 90 years.

Speaker 4 The IRC helps refugees whose lives are disrupted by conflict and disaster, supporting recovery efforts and responding within 72 hours of crisis. Donate today by visiting rescue.org slash rebuild.

Speaker 3 You know, you made the point earlier that it's enormously expensive to develop AI.

Speaker 3 I mean, the talent is high-priced, and it takes tons and tons of computing power to develop the systems, to run them once you have them, which means not a couple, three million dollars, but hundreds of millions in some cases or more, which means that the big companies in tech, Microsoft, Google, Meta, we all know the names,

Speaker 3 have an edge, but it's interesting, as I read your story,

Speaker 3 that's no guarantee of success, is it? Sometimes it's kind of an obstacle, having a big organization.

Speaker 2 Aaron Powell, you know, it's interesting. Let's use the example of Google.
Let's give Google credit first. They were so far ahead of almost everyone else on AI.
They hired some of the best talent.

Speaker 2 They were employing machine learning, deep learning long before most everyone else.

Speaker 2 You know, they did some of the more cutting-edge things. In fact, the breakthrough that led to ChatGPT was actually

Speaker 2 out of Google. Google had inside the company in around 2020 a chat GPT equivalent.
But

Speaker 2 Google takes in a lot of revenue. There's a lot of risk if this chatbot misbehaves.
There was famously this example of Microsoft, I think it was 2016, 2017, came out with Tay.

Speaker 2 And, you know, it was trained on social media and that kind of thing. And within 24 hours, it was a Holocaust-denying white supremacist.

Speaker 2 And of course, Microsoft, worrying about the reputational risk, pulled the plug on that rather quickly. And I feel like that's haunted the giant.

Speaker 2 So even though Google was far ahead, even though Google could have had their version of ChatGPT and it was Google that changed the world,

Speaker 2 they were scared of it. And, you know, never underestimate the ability of a giant to stumble over its own feet.
They have layers and layers of bureaucracy.

Speaker 2 They have, you know, huge public relations department that's whispering the CEO's ear.

Speaker 2 You know, I don't think it's a coincidence that OpenAI, a startup founded in 2015 was the one that set off the starter's pistol on this because they didn't have as much as at stake.

Speaker 2 They can afford reputation-wise to release ChatGPT. They could just make the decision without 10 layers of decision-making before they did it.
And so, yes, they have an advantage, but

Speaker 2 Google also has like... $100 billion of reserves, you know, where OpenAI has to go out and raise funds.

Speaker 2 they've raised roughly, I don't know, 20 billion so far, and there's talk that they've raised another 30 billion, and those, I might even be underestimating, and so that's 50 billion or so.

Speaker 2 You know, Google, they just pay for themselves, Microsoft, Meta, they all have deep, deep, deep reserves of money. And so

Speaker 2 it's almost like a race of attrition. You know, you can use these chat bots for free.
If you want the leading edge, cutting edge, you know, you have to pay, consumer would pay 20 bucks a month for it.

Speaker 2 But you know, most people are using these things for free and it's costing the companies a lot more than $20 a month to handle the heavy usage.

Speaker 2 And so these things are going to become more of a commodity. You know, there's a leapfrogging going on.

Speaker 2 Like, yes, GPT-4, that's OpenAIs, you know, when it came out, it was cutting edge, but then Anthropics Claude leapfrogged over that, and then others leapfrogged over that.

Speaker 2 And so they're all more or less as powerful, as

Speaker 2 useful as the other. And it's not clear how any of these companies are going to make money.
Google can afford to lose money on these things for five years plus. A startup, that's harder to do.

Speaker 3 Trevor Burrus, Jr.: Right, right. And so a lot of times you see the big companies buying smaller startups that have shown promise.

Speaker 3 It's interesting that this company called Open AI kind of became the public face of artificial intelligence in a way. It was a startup that didn't have the power of a Microsoft or a Google behind it.

Speaker 3 It was this guy, Sam Altman, and some other folks.

Speaker 2 Elon Musk.

Speaker 3 Yeah, Elon Musk, among others, right, right.

Speaker 3 And there's a moment that was sort of a critical transformational point when they released this version of ChatGPT, but that was preceded by a dinner at Bill Gates' house, which you described, which the house being as absolutely as magnificent as you would expect a Bill Gates' house to be.

Speaker 3 Tell us about that evening. What happened?

Speaker 2 So Microsoft, starting in 2019, started investing

Speaker 2 in Open AI.

Speaker 2 And so, you know, they had a financial stake. So Open AI would give Bill Gates, others at Microsoft, an early peek at what they were learning.
And Gates, who to him, AI is the holy grail of computing.

Speaker 2 He's been programming since before he was born, practically. And to artificial intelligence, the holy grail.
And he was impressed with, I think think it was GPT-3 or whatever,

Speaker 2 the most recent one he had seen. But he gave a challenge.
He said, I'm going to be impressed if it could

Speaker 2 ace the biology AP test. And he chose that one because it's not just regurgitating facts.
You need to analyze, you need to synthesize.

Speaker 2 You really have to show some sense of understanding and intelligence. And he thought that would be a great challenge.
And so, you know,

Speaker 2 he threw down the gauntlet and thought, dunk Alia from them in a few years, whatever. And

Speaker 2 not that many months later,

Speaker 2 he heard from OpenAI, okay, we're ready. And so in September of 2022,

Speaker 2 Gates hosted at his house

Speaker 2 a demo,

Speaker 2 OpenAI, and it was whatever, 30 people from Microsoft, from OpenAI,

Speaker 2 while someone was at a computer, the big screen was set up, and

Speaker 2 watching this computer take

Speaker 2 this test.

Speaker 2 And, you know, within two or three answers, people were just blown away. And in fact, it did get five out of five on the test.
It did pass the test.

Speaker 2 And that's when Gates became a true, true, true believer.

Speaker 2 You know, I thought I was, you know, in his mind, as he said, you know, I thought I was throwing down a gauntlet that would be a while, and suddenly, you know, it matched

Speaker 2 my expectations. In fact, and then they kept on playing with it and they would just ask it, you know, what would you say to a father, you know, worried about the health of his son?

Speaker 2 And it just kind of spit out an answer in his gates foot. It's like, it's kind of better answer than most of us could have given sitting around that room.

Speaker 2 And, you know, they just started playing with it. Gates started playing with it.
Others started playing with it. And it just blew them away.

Speaker 3 We're going to take another break here. Let me reintroduce you.
We are speaking with Gary Rivlin. He's a veteran investigative reporter.

Speaker 3 His new book is AI Valley, Microsoft, Google, and the trillion dollar race to cash in on artificial intelligence. We'll be back to talk more in just a moment.
This is Fresh Air.

Speaker 1 This message comes from BetterHelp. President Fernando Madera describes how BetterHelp Online Therapy has helped him.

Speaker 6 For me, sometimes I just need to to go and talk to somebody that is not going to judge me, right? It's going to be there and going to listen to me.

Speaker 6 And I can't start just saying, look, I'm not feeling right today. And it feels natural.
I love it.

Speaker 1 To get matched with a therapist, visit betterhelp.com slash NPR for 10% off your first month.

Speaker 4 Support for NPR and the following message come from Green Chef. Green Chef makes it easy to spend less time in the kitchen and more time enjoying fall.

Speaker 4 Green Chef supports your wellness journey with organic and seasonal produce recipes and weekly meal plans. Make this fall your healthiest yet with Green Chef.

Speaker 4 Head to greenschef.com slash 50fresh and use code 50fresh to get 50% off your first month, then 20% off for two months with free shipping. That's code 50fresh at greenschef.com slash 50fresh.

Speaker 3 You know, a few weeks ago, there was this development which kind of shook the stock market.

Speaker 3 This Chinese company called DeepSeek announced that they had created this artificial intelligence system at far less cost without the sophisticated microchips that American companies were using.

Speaker 3 It made Americans wonder, heavens, are we about to be overtaken?

Speaker 2 I don't know.

Speaker 3 Where does all this leave us? How important is this development?

Speaker 2 Right. So, I mean, to me, some of that was overstated.

Speaker 2 You know, Silicon Valley companies were experimenting with smaller models that required less compute

Speaker 2 power. You know, DeepSeek itself was venture-funded.
You know, it was cheaper, but hardly cheap.

Speaker 2 You know, they still, you know, cost millions millions to train it, presumably, and it costs millions, tens of millions to operate it. It just didn't require as much.

Speaker 2 And that really kind of was almost an existential threat

Speaker 2 to Silicon Valley, which they had put all their money, these tens of billions, hundreds of billions of dollars, into building ever bigger models that presume that you need ever more computer

Speaker 2 power. But, you know, a couple of things.
One, I think all it means is that

Speaker 2 instead of like, hey, we can do this at one-tenth the power, one-tenth the cost, I think they're just going to build 10 times more powerful models because they could do

Speaker 2 more with more with less.

Speaker 3 Do you mean the DeepSeek, the Chinese, or do you mean who?

Speaker 2 No, the American companies.

Speaker 2 They're learning from this. They'll integrate it.
Like I said, I feel like the AI companies I was following, they were already for a year plus paying attention to smaller models.

Speaker 2 Like, maybe you don't need this whole huge system to answer a simple question.

Speaker 2 You know, maybe we should have a bunch of smaller models and, like, okay, this one's an expert in this, that one's an expert in that. And we just have a smaller model give questions.
But, you know,

Speaker 2 I think what an OpenAI would say, other than the fact, an ironic statement that

Speaker 2 you used our model to train, I say it's ironic because OpenAI is being sued for

Speaker 2 taking the copyright, the intellectual property of the New New York Times, of bookwriters, of artists, of musicians, and all.

Speaker 2 But, you know, I think what's interesting about DeepSeek is it really gives hope to startups. Like, wait, okay, maybe you don't need as much money as we thought you do to create a company.

Speaker 2 But, you know,

Speaker 2 I do think it's important to understand that they still were using a lot of computer power.

Speaker 2 They still required a lot of money, just not as much as some of these larger companies that we've been talking about.

Speaker 3 Aaron Powell, you know, know, Reid Hoffman, the investor who's been very active in this area, is ultimately very optimistic about where AI is going to take us.

Speaker 2 Where are you on that scale?

Speaker 2 I do feel that AI is going to bring about incredible things. I think it's being overstated.

Speaker 2 You hear people say that it's going to close the divide between the developing world and the developed world. I don't think that's so.

Speaker 2 But there's this interesting study that came out recently, the idea of an AI tutor, a tutor in the pocket, that everyone has access, five billion people around the globe have a smartphone, and you can use that smartphone as a tutor.

Speaker 2 And so there was a study in Africa that like, let's let these kids after school have access to these AI tutors. And in six weeks, they showed two years worth of advancements.

Speaker 2 And I really do think around education, around science, you know, science is balkanized, right?

Speaker 2 It's, you know, it's specialties and subspecialties, and there's their own vocabulary lingo in every subspecialty.

Speaker 2 These large language models could read across specialties and connect the dots. They can make connections that no human being can do.

Speaker 2 And I think we're going to see some amazing scientific advancements. Creation of vaccines, of better therapies.

Speaker 2 There are some who predict, and I actually think there's a lot to it, that the mortality rate for most cancers are going to go way down because of AI.

Speaker 2 So I really do think AI could do some amazing things. It's just, I just don't know how bad the bad is going to be.

Speaker 2 If I had one wish, I wish we were dealing with the concerns that are within the line of sight, the stuff that we can imagine. Like, wait,

Speaker 2 it could be used for scams, it could be used in warfare.

Speaker 2 Instead of like this idea of the robots are going to take over and subjugate humanity, I guess that's possible, but not in the short term, not in the medium term, you know, just in kind of in the long term.

Speaker 2 And if we're deliberate about it, I think there's no doubt that AI could be a positive. You know, again, I just

Speaker 2 compare it to the internet. You know, is the internet a great thing?

Speaker 2 Like, no, I can tell you a lot of negatives with the internet, but you know, I think the internet has changed society in a lot of ways that, you know, we like.

Speaker 2 You know, the smartphone, the same kind of thing. So it's going to be a mixed bag.
And I guess I'm keeping my fingers crossed that

Speaker 2 despite the next four years, there's not going to be much regulation, not much checks and balances, that

Speaker 2 AI is going to be a net positive.

Speaker 3 Aaron Powell, speaking of guardrails,

Speaker 3 what rules, if any, do you have for your kids and their use of chatbots?

Speaker 2 You know, right after ChatGPT came out,

Speaker 2 middle school where my younger son goes, they kind of had this idea of banning it. And it's like, wait, wait, wait.

Speaker 2 They need to learn how to use this. I'll go back to what I was saying before that, you know, we have to learn how to use this.
What is this good for? And what are ways we can't rely on it right now?

Speaker 2 So, you know,

Speaker 2 one of my sons writes a composition, you know, like throw it into ChatGPT and get some feedback

Speaker 2 on it. Like,

Speaker 2 I may or may have not caught my older son,

Speaker 2 you know, using it to write an English paper. You know, within three sentences.

Speaker 3 Just told about a million people what you may or may not have done. Caught him doing.

Speaker 2 You know, within three sentences, it was obvious, like, okay,

Speaker 2 this is too perfect. This sounds like, you know, cliff notes for those of us who are old enough to know what a cliff notes are.
But it's like, go rewrite.

Speaker 2 So, you know, don't use it to write, but you use it as a research assistant. You know, use it for feedback.
And in fact, I see with one of my sons, you know, a teacher like, yeah.

Speaker 2 If you're writing something for science, use it and get some feedback on, you know, saying more clearly what it is.

Speaker 2 But, you know, it's a very personal choice, but I'm convinced that my kids, their life is going to be as dramatically different

Speaker 2 as mine was growing up before the internet and before mobile phones became pervasive. I really do think AI like the internet, like the phone within, you know, I'll say 10 or 15 years.

Speaker 2 I could be wrong on that, but at some point in the future is going to be at the center of their lives. And I think this next generation should get used to it because it's going to be critical to

Speaker 2 what they do, how they relate to the world, how they get employment. Aaron Powell,

Speaker 3 the company Inflection, that you write about, they had this chatbot, Pi.

Speaker 3 You had an interesting exchange with that chatbot about a medical issue your son had. Do you want to share that with us?

Speaker 2 Yeah, so we were facing this health crisis

Speaker 2 just as Pi was coming out. And usually what a reporter does when the chatbot comes out is they try to mess with it.
They try to get it to misbehave. They try to get it to jump the fence.

Speaker 2 But let me try dealing with this in a more authentic way. And, you know, I was really impressed.
You know, it had just the right tone, said all the right things, if not a little too perfectly.

Speaker 2 You know, it asked the right questions to get a dialogue going, you know, kind of in the fashion of a friend. Like, how is your son taking the news? How's the school handling it?

Speaker 2 How are you taking care of yourself through these stressful times? You know, it's a slew of questions, probably too many questions, but you know, it really picked up on nuance. It got little jokes.

Speaker 2 I told a funny moment from the sit-down with the neurosurgeon, you know, and it just responded like, you know, teenagers, am I right? You know, it gave me a lot of things to think about.

Speaker 2 But what was so interesting to me is that it also didn't mean anything to me. You know, there's this quote I love from an MIT sociologist, Sherry Turkle.

Speaker 2 You know, the performance of empathy is not empathy. You know, about expressing empathy, it's not really empathy.

Speaker 2 It's just algorithms parsing human language patterns, trying to like, oh, here's the right thing to ask and stuff. But you know,

Speaker 2 it really was an interesting experience. And I can understand like, you know, if people were lonely, if people didn't have, you know, a network of people to speak with, this could be really something.

Speaker 2 I think something people have to get used to is dropping this idea like, oh my God, you're going to have a friendship with a bot. You know, you're going to treat it like a therapist.

Speaker 2 Yes, of course, you should go to a licensed therapist to deal with your issues, but like, you know, what if you don't have a few hundred dollars or whatever uh it costs for a therapist every week and you know it's like they really do help you think through at least this bot pie really helps you think through one of the questions you should be asking yourself and also you know it was a really it was really interesting experience for me to really just try to feel like just your average user what they would feel like you know discussing something difficult brain surgery in this case of my son which by the way way, I should say, it was a very happy ending.

Speaker 2 Everything turned out fantastic. It's easy to talk about because of

Speaker 2 that. Good.

Speaker 3 I'm glad you mentioned that.

Speaker 2 But, you know, the bot gave me some interesting things to think about.

Speaker 3 Well, Gary Rivlin, thanks so much for speaking with us again.

Speaker 2 Oh, my pleasure. Thank you so much.

Speaker 3 Gary Rivlin is a veteran investigative reporter. His new book is AI Valley, Microsoft, Google, and the Trillion Dollar Race to Cash In on Artificial Intelligence.
This is Fresh Air.

Speaker 1 This message comes from Dell Technologies.

Speaker 1 It's time for Black Friday, Dell Technologies' biggest sale of the year for gift-worthy accessories and select PCs like the Dell 16 Plus featuring Intel Core ultraprocessors.

Speaker 1 Shop now at dell.com slash deals.

Speaker 4 Support for this podcast and the following message come from AmeriPrize Financial. Chief Market Strategist Anthony Saglan Benny shares the importance of a goal-based investment strategy.

Speaker 2 You have to know where you're going, right?

Speaker 3 What's the goal? What's the destination? By identifying those goals, you can construct a well-diversified portfolio that hopefully helps meet those goals.

Speaker 4 For more information and important disclosures, visit AmeriPrise.com/slash advice. Ameriprise Financial cannot guarantee future financial results.

Speaker 4 Securities offered by Ameriprise Financial Services, LLC, member of FINRA, and SIPC.

Speaker 1 This message comes from Capital One. Banking with Capital One helps you keep more money in your wallet with no fees or minimums on checking accounts.
What's in your wallet? Terms apply.

Speaker 1 See capital1.com slash bank for details. Capital One NA, member FDIC.

Speaker 3 Karen Russell's first novel, Swamp Landia, came out in 2011 and was a finalist for the Pulitzer Prize.

Speaker 3 Our book critic Maureen Corgan says she expects Russell's new novel, The Antidote, will be on a lot of prize lists this year. Here's her review.

Speaker 7 No one summons up the old weird America in fiction like Karen Russell does.

Speaker 7 Her tall tales of alligator wrestlers in Florida, homesteaders on the Gothic Great Plains, and female prospectors digging for gold mash up history with the macabre in a cracker barrel aged with dry humor.

Speaker 7 Russell's celebrated debut novel, Swamp Landia, came out in 2011.

Speaker 7 Since then, she's published a couple of excellent short story collections, but the wait for another novel was growing a little strained.

Speaker 7 I even heard speculation that maybe all the acclaim Russell received for her first novel had blocked her. Well, the antidote has just come out, and now we know why it took so long.

Speaker 7 American epics take a while.

Speaker 7 The antidote is set in a Dust Bowl-era Nebraska town called Uz, but it also reaches back to the earlier pioneer era Russell evoked in her short story masterpiece, Proving Up, which was made into an opera.

Speaker 7 The novel is framed by two true weather catastrophes the Black Sunday dust storm on April 14, 1935, in which people were suffocated by a moving black wall of dust, and a month later the Republican River flood, when twenty four inches of rain fell within one day.

Speaker 7 Much of what occurs between those two disasters is also true emotionally, but in Russell's worldview, the fantastic and the familiar coexist on the same plane.

Speaker 7 Our central character here is a prairie witch who goes by the name the antidote.

Speaker 7 Part huckster, mostly healer, she, like other prairie witches, promises to treat what ails her customers by taking away whatever they can't stand to know, the memories that make them chase impossible dreams, that make them sick with regret and grief, whatever cargo unbalances the cart.

Speaker 7 I can hold on to anything for anyone. Milk, honey, rain water, venom, blood, pour it all into me.
I am the empty bottle.

Speaker 7 Lying in a trance, the antidote absorbs the heaviness, but not the details of her customers' stories, which they sometimes want back.

Speaker 7 After the Black Sunday dust settles, however, the antidote is horrified to realize she feels lighter, vacant, some awful force has robbed her of the stories she safeguarded.

Speaker 7 Who knows how her more violent customers will react when they discover they can't make withdrawals?

Speaker 7 Other narrators step in to amplify Russell's peculiar vision of life in Uz.

Speaker 7 There's Del Oletsky, a teenage girl whose single mother was allegedly murdered by the lucky rabbit's foot killer, so-called because he leaves a bloody rabbit's foot near his victim's bodies.

Speaker 7 Del lives with her uncle Harp, whose farm is mysteriously untouched by the all-enveloping dust. A federal agency photographer, a black woman named Cleo Alfrey, eventually turns up in Uz.

Speaker 7 Cleo explains her work by saying she's making advertisements for Roosevelt's New Deal programs. She's also painfully aware of whose faces carried the most weight with Congress.

Speaker 7 Actual Depression Era photographs are scattered throughout this novel, but the camera Cleo depends on goes Twilight Zone Haywire, photographing the past and possible futures of the town and surrounding terrain.

Speaker 7 Like Cleo's camera, Russell's instrument, her language is uncanny.

Speaker 7 Swathes of the spellbinding final third of this novel move deeply into the past, specifically into the buried memory of how Harp Oletsky's parents in Poland grabbed at the offer of free land in Nebraska, land they come to realize that was occupied before their arrival.

Speaker 7 Here's Harp's father guiltily recalling how he made peace not only with that land grab, but with racial hierarchy in America.

Speaker 7 I was born a serf in all but name. My skin is the color of an unwashed onion.
In America, this placed me ahead of many, on a low rung of the ladder, but higher than the black porter.

Speaker 7 I heard the ticking pulse of a sick relief. Not me, not me, not me.
The same feeling I once had whenever one of my brothers was chosen over me for a beating.

Speaker 7 In the antidote, Karen Russell, America's own prairie witch of a writer, exhumes memories out of the collective national unconscious and invites us to see our history in full.

Speaker 7 There are, alas, no antidotes for history.

Speaker 7 Our consolations are found in writers like Russell who refract horror and wonder through their own strange looking glass, leaving us energized for that next astounding thing.

Speaker 3 Maureen Corrigan is a professor of literature at Georgetown University. She reviewed The Antidote by Karen Russell.

Speaker 3 On Tomorrow's Show, New Yorker staff writer Andrew Morant joins us to discuss how podcasts, live streams, and YouTube channels have become the platforms where men who feel disillusioned and alienated go to feel seen and heard, many of them gravitating toward the MAGA movement.

Speaker 3 I hope you can join us. To keep up with what's on the show and get highlights of our interviews, follow us on Instagram at NPRFreshAir.

Speaker 3 Fresh Air's executive producer is Danny Miller. Our technical director and engineer is Audrey Bentham with additional engineering support from Al Banks.
Our managing producer is Sam Brigger.

Speaker 3 Our interviews and reviews are produced and edited by Phyllis Myers, Anne-Marie Baldonado, Lauren Krenzel, Teresa Madden, Monique Nazareth, Thea Challener, Susan Yakundi, and Anna Bauman.

Speaker 3 Our digital media producer is Molly C. V.
Nesper. Roberta Shorrock directs the show.
For Terry Gross and Tanya Mosley, I'm Dave Davies.

Speaker 1 This message comes from the Council for Interior Design Qualification.

Speaker 1 Interior Designer and CIDQ President Siavash Madani discusses why certified professionals know that good design is more than just how something looks.

Speaker 5 Being NCIDQ certified means you've qualified to protect the health, safety, and welfare of the public in the spaces that you design. Good design is never just about aesthetics.

Speaker 5 It's about intention, safety, and impact. So an NCIDQ certified interior designer must complete a minimum of six years of specialized education and work experience and pass the three-part NCIDQ exam.

Speaker 5 All three exams emphasize and focus on health, safety, and welfare of the occupants.

Speaker 5 Being NCIDQ certified means that you've proven your knowledge and skills through rigorous exams and are recognized as a qualified interior design professional.

Speaker 1 Learn more at cidq.org/slash npr.