The new wave of AI tools is changing us

Make Me Smart April 11, 2023 Transcript

Note: Marketplace podcasts are meant to be heard, with emphasis, tone and audio elements a transcript can’t capture. Transcripts are generated using a combination of automated software and human transcribers, and may contain errors. Please check the corresponding audio before quoting it.

Kimberly Adams 

Hello, I’m Kimberly Adams. Welcome to make me smart, where none of us is as smart as all of us.

Kai Ryssdal 

I’m Kai Ryssdal. Tuesday 11 April is what calendar says. Good to have you all along. So we are going to do a little AI today. We’re gonna talk about the buzz we’re gonna talk about how worried you ought be or maybe not. We’re gonna talk about a lot of stuff.

Kimberly Adams 

Yeah, in particular, we we want to know about the cultural impact some of these AI tools might be having already, how they’re changing our relationship with technology, maybe what’s coming down the road. So here to make us smart about all of this is Kyle Chayka, staff writer at The New Yorker covering, as appropriate, technology and culture. Welcome to the show.

Kyle Chayka 

Thanks for having me.

Kimberly Adams 

I know that chat GPT and versions of this technology have been around for a couple of years now. But it feels like it’s been just in the last couple of months, where it really entered the mainstream. And we’ve talked on the show about whether or not this is like the printing press or the advent of the Internet or the rollout of the iPhone. What would you compare it to?

Kyle Chayka 

It’s a really hard question. But I do think it rises to that level of the printing press or the internet, where it’s this tool that like fundamentally shapes everything we do, how we think, how we interact with the world. So I kind of see it influencing everything that happens going forward.

Kai Ryssdal 

Here’s the thing, though, you know, the printing press for as good as Mr. Gutenberg was, you know, getting that Bible done, it took a long, long time for that to roll out. And here we are now are now, I mean, almost literally at the speed of light, right? I mean, that’s how fast this thing works. And, and we’re having a tough time wrapping our brains around it. Right?

Kyle Chayka 

Yeah, it’s just happening so fast. I mean, as you said, even in the past few weeks, things have changed. There are more new tools, the tools are smarter, it’s just impossible to keep up.

Kimberly Adams 

So what, at this point, do these chatbots like ChatGPT and Bing Ai do well?

Kyle Chayka 

I think the best things that they do are kind of office assistant tasks. Like they can look up facts for you, they can generate an itinerary, they can summarize the report or write a report for you. And they can kind of suggest things for you to write that a human still certainly has to edit and think through. But yeah, I think these kinds of like office tasks, computer tasks, writing, emails, communication, all the things that kind of consist of, that our jobs consist of, that are being influenced by them.

Kimberly Adams 

Don’t panic

Kyle Chayka 

I feel a little panicky, honestly,

Kai Ryssdal 

Well say more about that. Keep going.

Kyle Chayka 

Yeah, I mean, I think the fluency with which chat GPT and other tools can generate text, as a writer that’s a little scary to me, because, you know, I make a living being a human being who writes things, and hopefully writes them well and has original insights. But for like the vast majority of writing that’s produced in the world, it’s possible that these tools can just do it faster, if not better.

Kai Ryssdal 

Can they do it tomorrow? Or is it going to be a year? Or is it going to be 10 years? It’s not gonna be 10 years nevermind?

Kyle Chayka 

No, I feel like it’s six months to a year. I mean, I talked to people now who are using these tools actively, who are using using image generating tools and art therapy, using text generating tools to do their pitch decks. Like it’s happening right now.

Kimberly Adams 

You know, I wanted to ask you about image generation. And, you know, for our purposes, audio generation. You know, we’re already seeing these very convincing photos of political leaders. I saw on, I don’t know if it’s an Instagram account, or TikTok account that basically converts many Republican members of Congress into very realistic images of people who dress in drag. And then we all saw the pope in the coat that was not real, and images of Trump being arrested. And you know, I was just reading a story in The Post today about AI being used to make extraordinarily convincing pornographic images. And I wonder how you think that is changing things and how quickly?

Kyle Chayka 

Yeah, I mean, it clearly is changing things right now. I mean, people are being fooled by these disinformation images. Like I was completely fooled by the pope wearing a very fancy jacket. I totally thought it was real. I mean, I think not every AI generated image or piece of media is that convincing? Like, there are also fake videos of Biden and Trump playing Minecraft, which are like, very funny, but definitely not real in any way. I think I saw them compared to political cartoons. So I feel like the majority of this stuff, I don’t think is meant to fool people necessarily, or at least doesn’t fool people right now. But it is this kind of new creative form and people are playing around with it. Like I listened to this track by a French rap duo that just used Jay Z’s voice as a rapper in their song. And that just totally blew my mind.

Kai Ryssdal 

So how should we think about this? Because panic gets you nowhere but obviously, we can’t ignore it. What’s the happy medium?

Kyle Chayka 

I think insofar as it’s possible, I’m trying to use the tools and experiment with them and kind of feel the limits of what they can do, if even just to convince myself not to worry too much. Like chat GPT can’t generate a good 1200 word article, which is kind of my bread and butter I would say. But, you know, it’s possible… I want to know how people are using these day to day and incorporating them into their lives. And so I think seeing what happens will make us understand better if this is just like the new word, document or email or if it’s some kind of more insidious change.

Kimberly Adams 

What have you seen so far as you’ve been observing other people working with the technology and yourself working with the technology, companies working with the technology? How is it changing us?

Kyle Chayka 

I mean, I think it might change our minds faster than what we do. Maybe that doesn’t, that’s not very clear. But it’s, you know, it changes our mental imagination of what we see in front of us. So when when I look at an image online now, or I maybe read an email or look at a PDF, there is this like niggling feeling that it might be AI generated. So it makes me almost doubt the authenticity of the things in front of me in a way that I absolutely didn’t a year ago. Like, you know, you see a photo of the Pope wearing a leather jacket or in front of a crowd. Maybe it’s real, maybe it’s not. The rise of these tools means we don’t know in that first instant.

Kai Ryssdal 

What’s interesting about that is that the skepticism about the Pope is an interesting pop culture thing. But there’s real disinformation and misinformation going on in this society right now and as we all know, in our politics, and maybe a little initial hesitation, is this real would be good?

Kyle Chayka 

Yeah, couldn’t build some more skepticism, or maybe fluency in what we see online. But that almost feels like a losing bet given our past 10 years of history with the internet. Like trusting people online to think twice has not been a good bet I would say

Kai Ryssdal 

It’s not a money winner. Yeah.

Kimberly Adams 

Well, and I’m remembering in the early days of the January 6 hearings, you know, Benny Thompson said, he quoted song lyrics said, “Are you going to believe me or your lying eyes?” Right? We rely so much on visual evidence of things for proof, on documents for proof. And I wonder what it will mean, especially in our political climate, when you can’t trust those things so easily anymore.

Kyle Chayka 

I mean, you would hope that the AI companies are building in safety mechanisms like, you know, stamps or watermarks, that that show that something is AI generated. And there are tools that allow you to put in text and kind of give a good judgement of if it’s AI generated or not. But I think I mean, with the rate that these things are improving, that’s going to be harder and harder to track. And the tools themselves are getting more and more accessible. Like you saw the Facebook AI model leaked online just in the past few weeks.

Kai Ryssdal 

So so all in all, are you hopeful or fearful about this technology and what it’s going to mean for, you know, all of humanity?

Kyle Chayka 

I think I’m fearful in the short run and maybe hopeful in the long run, ideally. I think short term, we’re going to have to adjust a lot of our behaviors, particularly in regards to digital content. We’re gonna have to see how this impacts our livelihoods and how it changes work for a lot of people. In the long term, you know, I hope that it shows that maybe human labor isn’t so irreplaceable after all, and we should have universal basic income and a vast tax on AI tools and training data. But you know, that’s the kind of utopian future.

Kimberly Adams 

What then would you say?, and in terms of like, how we make this adjustment. Because I know a lot of people who are aware it’s out there, but they feel like they’re so behind the curve already that just entering the space feels daunting. Like, if you were to give somebody sort of an easy thing to hand off to AI that may help them kind of get a sense of it, and maybe feel a little bit more aware, if not comfortable, what would what would you suggest?

Kyle Chayka 

Yeah, I mean, I think the Bing search that’s driven by AI now is a pretty interesting tool. It’s not completely publicly available. But if you do chat with that bot, and ask it to generate a travel itinerary, list you 10 vegan recipes or something like that, it is really fascinating to see it spit back that list immediately or write something, write a piece of content that that is specifically tailored to what you needed. And then the image generators too. I think playing around with them gives you a sense of what they can and can’t do. And maybe it gives you a better sense of what their flaws are. So if you can go into Midjourney’s discord, and I don’t know, generate a funny image of a dog or something, you might… it’s fun, it’s completely fun and interesting. And it gives you more fluency in what’s actually happening. So I think those are useful exercises.

Kimberly Adams 

Kyle Chayka is a staff writer at The New Yorker covering technology and culture and trying to keep us all from panicking and running around screaming. Thank you so much.

Kyle Chayka 

Thank you! Not replaced by a robot yet.

Kimberly Adams 

Not replaced by a robot yet. That’s the motto. I got replaced by Siri ages ago. So I’m alright.

Kyle Chayka 

Thank you for having me

Kai Ryssdal 

Yeah, I want to be hopeful. I’m a little I’m a little weary right now. I would say that’s where I am. I think that’s where I am.

Kimberly Adams 

I mean, I just I have a lot of faith in human resilience. And we’ve survived some pretty terrible inventions over the millennia… good and bad, and pretty much everything that’s started out as being something for good has been manipulated into bad and we’re still here for now. So I’m gonna choose to have faith in us. Fair enough. Fair enough. If you have used ChatGPT or you know, any of the other AI devices that are out there, which one do you like? Which one you don’t? Let us know. Let us know how it’s useful to you. Or you know, perhaps it’s not, I don’t know. Our number is 508-827-6278. 508-U-B-SMART. Email us if you’d like. make mesmart@marketplace.org We are coming right back.

Kai Ryssdal 

All right news. That’s where we are. That’s how you know we’re doing the news is that little thing you just heard. Kimberly go. It’s called a sting technically in radio. But come on.

Kimberly Adams 

Look, just following up on the conversation we just had, there are so many different camps of where to fall as we’re thinking about where we’re going with AI, what role it’s going to have in our society. And the Washington Post had this really interesting piece, sort of breaking down some of these different camps, including, I’m seeing how they categorize them, the AI safety camp, the AGI believers, artificial general intelligence, the AI Doomers, the AI ethicist. And it breaks down sort of who’s in each of these groups, where they stand on where AI is going, and how much power and influence they have. And what I thought was really valuable about this breakdown is that there are really smart people in all of these different groups. And so if folks like you and I, or people in the audience are kind of like, we don’t know where to where to land with this, you’re not alone. Some of the brightest minds in the country and in the world are also still trying to figure out where we’re going with all this.

Kai Ryssdal 

That’s a really hard problem, right? It’s just, it’s really, really difficult. And I’ll just, I’ll just, I know, you’ve got another news item you want to get to, but I’m just gonna jump right in here. Because because mine is sort of aligned. Just… so AI has been a thing now for a while in this society, right? We’ve all had, you know, our experiences with it, but ChatGPT and the real maturation of it seems to have come out of nowhere, like in the last two, three months. And so now the federal government, courtesy of an organization inside the Commerce Department called the National Telecommunications and Information Administration, NTIA for short, is just now taking the first steps toward trying to figure out what regulation ought to look like, if we decide we need to regulate this stuff, and is seeking public comment. So I just two thoughts. Number one is a little bit like hackers, right? The good guys, which is to say, you know, the people in charge of keeping everything orderly, are always a step behind the people developing the technology, whether it’s AI or Blackhat hackers or what have you. And I think that’s going to be a real challenge. Because as we’ve seen, this has moved so so so fast. Also, I hope there are a lot of really smart people out there who are going to comment on this thing, because opening it up for public comment could be just you know, a big barrel of monkeys or snakes or whatever you want to call it.

Kimberly Adams 

Regulations.gov friends. Public comments are your friend.

Kai Ryssdal 

You are the biggest geek. Oh my god.

Kimberly Adams 

Well, since we’re going all tech today, you know how sometimes you have these like moments where you’re paranoid about technology and you’re like, eh probably overreacting, my phone isn’t really listening to me.

Kai Ryssdal 

Oh but it is. But it is. Explain all those ads that pop up on my Insta with me not having clicked on anything.

Kimberly Adams 

Exactly. So I had one of these moments today, when I came across this story. In the show notes we will have a link to the NBC news coverage of it, but it was everywhere. That the FBI is now warning people about using public phone charging stations. So it’s the place like in an airport or at a restaurant where the the outlet has like the USB plugs and you can plug directly into the USB and charge your phone. Because according to FBI Denver, “avoid using free charging stations at airports, hotels or shopping centers. Bad actors have figured out ways to use public USB ports to introduce malware and monitoring software onto devices. Carry your own charger and USB cord and use an electrical outlet instead.” Just because you’re paranoid doesn’t mean everyone’s not out to get you.

Kai Ryssdal 

That’s exactly right.

Kimberly Adams 

Yeah, and I mean, like, I’ve definitely had these moments where my battery’s been low and I’m just like, Oh, I know I’m not really supposed to plug into public, you know, charging but it’s right there and I don’t have the little plug that actually has the plug on it, I just have the USB cord. And sometimes I’ll be like in an Uber or Lyft and they offer you to charge your phone and I’m just like, is it a trap?

Kai Ryssdal 

This is now going to become a slide that we’re gonna have to spend six minutes on in our next IT security training. I guarantee it. Don’t pick up USBs in the parking lot and put it in your computer and don’t use public charging stations. I’m telling you

Kimberly Adams 

Yeah, I mean, it was one of those things that, you know, I kind of suspected, but I didn’t realize it was a real thing. And I guess there hasn’t, there haven’t been a ton of very public reports of people who’ve been hacked this way, because it’s a little bit hard to trace it. Because it kind of depends on when you figure out that your phone’s been compromised. But yeah, plug into the electrical outlet instead friends.

Kai Ryssdal 

You know it’s so interesting, sorry, I’m just thinking through the logistics, the actual, you know, signal path of this thing. So I always just assumed that those USB ports in like airports and whatever, were just somehow plugged into the power lines, and they just drew power from that, you know, minus whatever transistors or transformers or resistors would would step down the power. But this sort of implies that it’s something else.

Kimberly Adams 

So this is what I learned from reading in on this today is that there are kind of two different types of USB ports. Some only provide power for charging only, and some transfer power and data. And you know how sometimes you’ll plug your phone into, like your USB from charging your phone into a computer, and it’ll ask you, if you want to trust this device. That’s a signal that it has the option to send data back and forth. That’s why you can like upload, you know, from memory cards and stuff on one way to another. So some USB ports do data, some USB ports only do you know charging. And in the same way that people have like credit card skimmers on ATM machines or on gas stations, people are, you know, doing something similar when it comes to some of these USB charging ports. So… sigh. Gotta be vigilant, gotta be vigilant. Alright, that is it for the news fix. Let us move on to the mailbag.

Mailbag 

Hi Kai and Kimberly. This is Godfrey from San Francisco. Jessie from Charleston, South Carolina. And I have a follow up question. It has me thinking and feeling a lot of things.

Kai Ryssdal 

Alright we’ve been talking a bit about AI and Chad GPT not just today, but over the last number of weeks and months. Here is what one of y’all had to say.

Gayla 

Hello my favorite humans.The other day I suddenly had to write a letter of reference for one of our extra delightful theater volunteers and had zero minutes to dedicate to this important task. Enter chat GPT and voila a letter that expressed exactly what I wanted to say.  But here’s the kicker .. Next I typed in, tell me what you know about them make me smart podcast. The answer? A lengthy informative yet pretty darn inaccurate result, especially the lead sentence: ‘The Make Me Smart podcast is a popular podcast produced by American Public Media and hosted by Kai Ryssdal and Molly Wood.’ Say what? It goes on to cover things like your segment called ‘what’s making us happy.’ Still, the chat GPT tab may remain forever open on my laptop. Till next time. This is Gayla signing out from sunny Tucson, Arizona.

Kai Ryssdal 

Wow, how about that?

Kimberly Adams 

I mean, that’s not super wrong. I mean, you and Molly did start the show. And we are indeed sort of kind of produced by American Public Media in that it’s kind of sort of our parent company-ish? It’s a weird, it’s a weird hierarchical structure. But so it’s it’s accurate-ish. And I think that’s that’s the that’s the takeaway, right? It’s accurate enough to make it right.

Kai Ryssdal 

Yeah plausible. Right right right.

Kimberly Adams 

Which is what makes it so inherently risky. All right. Well, before we go, we’re gonna leave you with this week’s answer to the make me smart question, which is what is something you thought you knew, but later found out you were wrong about?

Alana 

This is Alana Ratliff for Wilmington, North Carolina. Something that I thought I knew but later found out I was wrong about was asparagus pee. I thought some people when they ate asparagus had stinky pee, and some did not. Later, I found how the No, everybody has stinky pee if they eat asparagus, it’s just some people can’t smell it. There’s actually been studies on this to determine the genetic variations that can be causing the asparagus as naamyaa I’m probably pronouncing that word wrong. But anyways, I’m smarter about my asparagus pee now. Have a great day.

Kai Ryssdal 

Something for everybody on this podcast, ladies and gentlemen.

Kimberly Adams 

Thanks I guess? I mean yeah. Learn something new everyday.

Kai Ryssdal 

All right, well, asparagus pee aside, what is something you thought you knew but later found out you were wrong about. Leave us a voice message with your answer to the make me smart question. Our number is 508-827-6278. 508-U-B-SMART. Man. Didn’t see that one coming.

Kimberly Adams 

Make Me Smart is produced by Courtney Bergsieker. Ellen Rolfes writes our newsletter. Our intern is Antonio Barreras. Today’s program was engineered by Juan Carlos Torrado with mixing by Mingxin Qiguan.

Kai Ryssdal 

Ben Tolliday and Daniel Ramirez composed our theme music. Our acting senior producer is Marissa Cabrera. Bridget Bodnar is the director of podcasts. Francesca Levy is the executive director of Digital and On Demand. They also work for her. And Marketplace’s Vice President and General Manager, for whom we all work, is Neal Scarbrough. It’s funny how some days there’s a lot of music leftover when we’re done with the credits and some days there isn’t.