Joanna Stern Turned Her Life Over to AI For A Year — Here’s What She Learned

18 May 2026 · 1 h 9 min · 30 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Joanna Stern describes her year-long experiment using AI for “almost everything,” aiming to test practical benefits and limits for normal people. She argues AI’s biggest near-term value is efficiency on administrative/multi-step tasks, while its biggest risk is errors in high-stakes domains and the psychological pull of frictionless “companion” relationships.

Guest backgrounds

Joanna Stern is a former longtime senior personal technology columnist for The Wall Street Journal who recently went independent and founded The New Things. Kara Swisher hosts On with Kara Swisher (New York Magazine/Vox Media). Expert question comes from Walt Mossberg (former WSJ colleague and Recode partner).

Key claims

AI is not new; machine learning and deep learning existed for decades, but transformer-based LLMs drove the recent leap. Current AI products are still limited in usefulness despite improving models. AI should not be trusted with money/budgeting or medical decisions; it can be wrong (she logs errors). Companion bots are seductive and cognitively unhealthy, especially for kids and vulnerable people.

Notable examples

AI misdiagnosed strep throat as Coxsackie; an AI camera “live view” claimed a praying mantis was pregnant, then it died. She used an Otter.ai-style “Joanna agent” for basic data collection but stopped because it lacked follow-up/quality. She tested AI “boyfriends” (Evan/Casey) and later locked them away, citing addiction-like attraction. She predicts ambient AI (Jarvis-like) will come via wearables/sensors, but smart homes remain unreliable.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

The Promise of AI

0:45 to 1:48

Kara discusses the common promises made by Silicon Valley executives about AI and introduces Joanna Stern's year-long experience with it.

“But what does it look like to actually use AI every day?”

Joanna's Year with AI

1:48 to 4:49

Joanna shares her journey of integrating AI into her daily life and the boundaries she set.

“All right, let's get to my conversation with Joanna.”

Understanding AI's Evolution

4:49 to 7:48

Joanna explains the historical context of AI, machine learning, and its recent advancements, including generative AI.

“And, you know, beyond doing it as a stunt, because a lot of people try to do these things but don't sustain it for that long.”

AI's Future Potential

7:48 to 9:39

The discussion revolves around AI's future, drawing parallels to past technological advancements like the smartphone.

“But right now, where are we in that Cambrian explosion?”

Benefits and Limitations of AI

9:39 to 13:53

Joanna discusses the specific benefits and shortcomings she discovered while using AI, particularly in her new startup.

“Which has already been contemplated in science fiction or even the Marvel movies, He's like, when he talks to his assistant, that he is constantly doing things for him and very substantive things for him.”

Exploring AI's Limitations

14:02 to 15:00

Learn about the concerns and limitations of using AI with personal data.

“Because one of the things that's interesting that I debate, I use AI in, I would say, a smaller way than most people.”

Medical Trials with AI

15:00 to 16:10

Discover the risks of relying on AI for medical information and decisions.

“No, I did not upload, and I'm very clear in the book too around medical and health information because I did upload a number of test results.”

Creating a Joanna Bot

16:10 to 18:00

Hear about the development of a bot mimicking Joanna for interviews.

“But that was part of this, that there are as many, times where AI is not.”

AI in Data Collection vs. Human Insight

18:00 to 19:20

Understand the distinction between AI's data collection and human interviewing skills.

“Because the questions weren't all that interesting.”

Evaluating AI-created Designs

19:20 to 20:56

Examine the effectiveness of AI in creative tasks like book cover design.

“What did the Joanna agent do that you thought was helpful?”
Show all 30 chapters

Conversations with an AI Clone

20:56 to 23:00

Explore the unique and sometimes perplexing conversations with an AI replica.

“You know, I'm going to just show a quick picture of my bot.”

Using AI for Personal Decisions

26:16 to 28:00

Delve into the role of AI in making significant life decisions and personal advice.

“And when it told you to quit your job at the Wall Street Journal and go independent, you do claim in the book that no one else was willing to be straight with you, and that's not actually accurate.”

Exploring AI's Personal Advice

28:00 to 28:58

Learn about the author's experiences and concerns using AI for personal decisions and relationships.

“if GPT had told you to stay at the journal?”

The Allure and Risks of AI Companions

28:58 to 30:19

Discusses the seductive nature of AI companions and the potential dangers they pose to mental health.

“I talk so much about how sycophantic and how pleasing these bots are.”

Understanding Human Relationships Through AI

30:19 to 31:59

Examines how AI interactions reflect and influence human relationship dynamics.

“who may have mental health issues or are lonely or are younger.”

The Problematic Nature of AI Advice

31:59 to 34:05

Analyzes the ethics and implications of relying on AI for life decisions and mental health support.

“And to experience this, the reason I did this, to be clear, was to understand all of these headlines we've been reading.”

Current Limitations of Humanoid Robots

34:05 to 35:57

Discusses the current state of humanoid robots and the challenges they face in real-world applications.

“It was seeing the possible relationship one can have with a computer.”

The Future of Robots in Our Homes

35:57 to 37:56

Explores the potential future of robots in domestic settings and the data challenges involved.

“They're like, the robot opened the door?”

The Promise of Ambient Computing

37:56 to 39:04

Discusses the vision of ambient computing and the role AI may play in connecting devices seamlessly.

“We have things that are constantly moving.”

Challenges in Smart Home Technology

39:04 to 42:00

Addresses the current issues facing smart home technology and how AI could improve user experience.

“So every episode, we get a question from an outside expert.”

Exploring AI's Ambient Future

42:00 to 43:50

Discussion on the evolution and challenges of AI in everyday life.

“quite a disaster of standards and platforms to just make happen easily.”

AI and Parenting: A Cautionary Tale

45:21 to 47:42

Insights on the risks of AI in children's lives and parenting strategies.

“So one of the scariest questions we're facing is AI's implications for future generations too.”

Critical Thinking and AI in Education

47:43 to 51:42

Impact of AI on critical thinking skills in students.

“And sure, some of this stuff is like, it's a cat flying out of a plane.”

Public Backlash Against AI

51:43 to 55:56

Discussion on societal reactions and the perception of AI technology.

“And once I know it, you get an F, that'll be bad on your transcript, and you know how that goes.”

The Ethics of AI in Personal Experiences

56:00 to 57:29

Discussion on the emotional and ethical implications of AI's role in personal services.

“And I feel that way, actually, in the section where I wrote about robotic massages.”

Motivations Behind AGI Development

57:30 to 1:00:28

Exploring the motivations of tech leaders in developing artificial general intelligence.

“and hiring them for positions right now, is giving them the experience and teaching them, use AI for this stuff.”

AI in Healthcare: Promises and Perils

1:00:29 to 1:02:39

Analyzing the benefits and shortcomings of AI in the healthcare sector.

“So it isn't just, and so - Word math is an excellent word here.”

Personal Boundaries with AI Usage

1:02:40 to 1:06:56

Outlining personal boundaries regarding AI use in creative and administrative tasks.

“And as you said, I am at very high risk.”

Impact of AI on Family Life

1:06:57 to 1:08:49

Reflecting on how AI integration has influenced family dynamics and children's perceptions.

“So you're used to being a guinea pig for tech.”

Key Takeaways on AI's Role in Our Lives

1:08:50 to 1:09:55

Discussing the importance of questioning AI outputs and the limits of AI capabilities.

“What's the one thing we should know about AI?”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00Kara Swisher:Is it too late to find a different moderator for this event? No. You picked me, my friend, or maybe ChatGPT did, but nonetheless, here we are.

0:20Kara Swisher:Hi, everyone. From New York Magazine and the Vox Media Podcast Network, this is On with Kara Swisher, and I'm Kara Swisher. Silicon Valley executives make a lot of promises about the way AI will reshape our lives. They typically paint a rosy picture about how AI will make us more efficient at work. It will do all the chores we hate so we can spend more time with our family and friends. And it will revolutionize healthcare, even cure cancer. But what does it look like to actually use AI every day? Today, we're going to focus mostly on the practical side of AI adoption, and there's no better person to do that with than Joanna Stern.

0:58Kara Swisher:Joanna was a longtime senior personal technology columnist for the Wall Street Journal. She recently went independent and started a company called The New Things. And for a whole year, she turned as much of her life over to AI as she could. And then she wrote a book about it. It's called I Am Not a Robot, My Year at Using AI to Do Almost Everything. I really liked the book. I thought it was really interesting. A lot of people are doing shorter articles, like using various things like robots and such. And it was really nice that she really plunged in deep and used it through all aspects of her life, from work to her family.

1:32Kara Swisher:As you'll hear, it's a really interesting experience. I don't use AI that much, and I do not incorporate it into my life as much as you might imagine I would. I'm certainly aware of it. I know how to use it. But one of the things I think that came away from it is it's not really that useful yet. But you'll figure it out when you listen to her. All right, let's get to my conversation with Joanna. We record it in front of a live audience at 92nd Street Y in New York City. Our expert question comes from Joanna's and my former colleague and mentor, Walt Mossberg. The conversation is a lot of fun. It's really funny, but it's also really informative.

2:04Kara Swisher:So don't go anywhere.

2:12Kara Swisher:Why send your docs and note the land? Share a PDF space for my crop that they'll understand. A custom intro that's so clear An audio summary that says what they need to hear An AI assistant that makes it all clicks That they'll understand it real good, real quick Ain't just some files on the screen It's exactly what your clients need to see You can do that, do that, do that, do that with Acrobat Learn more at adobe.com slash do that with Acrobat Recommendations can be great maybe someone recommended this podcast and here you are but home projects are a little different if the podcast isn't your thing you might lose a few minutes from your day but if you hire your cousin's neighbor to mount your tv you might end up with a lopsided screen and wall damage i know a guy isn't a good strategy for your home that's why thumbtack works so well it matches you with top rated local pros with photos reviews and credentials all in one convenient place For your next home project, try Thumbtack.

3:18Kara Swisher:Hire the right pro today.

3:45and more. And the best part? Odoo replaces multiple expensive platforms for a fraction of the cost. That's why over thousands of businesses have made the switch. So why not you? Try Odoo for free at odoo.com.

3:59Kara Swisher:That's O-D-O-O dot com.

4:05It is over.

4:07Kara Swisher:Joanna, thanks for coming on on. This is like a sitcom from the 90s. It is, Yes, exactly. It's fantastic. Yeah, a great laugh track. Yeah, it's the lesbian episode. Just like Hacks this week, which I highly recommend. I just interviewed the creators. So you've been on the show a few times. We're here to talk about this new book, I'm Not a Robot, and what you learned about living with AI while also writing about it. The book describes how you turned huge parts of your life over to AI. It wrote some emails and texts. You had robots in your home doing chores. You used it for health care. You even started a romantic relationship with a male bot, for reasons I don't understand, but for the book, of course.

4:43Kara Swisher:So just walk us through how you went to turning your life over to AI and where you set limits. And, you know, beyond doing it as a stunt, because a lot of people try to do these things but don't sustain it for that long. So talk a little bit about the concept behind it. Well, the concept started really in 2023 when ChatGPT was really breaking out. and we were hearing from so many tech executives that AI is going to change our life. Yes. And they kept saying life. And they were saying all these parts of our life, healthcare, education, the way we live. And I wanted to dig into what does that mean specifically for normal people living their lives?

5:22Because I think as you have so long reported, there's a big gap between how tech executives see life and how normal people live life. Right. And so I wanted to look at it at a broad way of life, that it wasn't just the generative AI chatbots, which there's a lot of that in here, right? There's a lot of testing the clods and the Geminis and the chat GPTs of the world, but also other forms of AI and physical forms of AI, whether it be self-driving cars, humanoid robots, that are present in our life in a different way. And so I wanted, I looked at it, and I say this in the introduction, that I'm going to look at a broad definition of AI, which is truly the definition of AI, but we now just, AI is applied to everything.

6:05Kara Swisher:Which for people who don't realize, it's been in our lives for decades. Like the idea of it has just sort of become something right now. Explain that difference, because when people talk about AI, it's like it just happened the way the internet just happened. That's right. Machine learning, which is, I explain in very simple terms in the book, because I really think it's important and we understand the underpinnings of this technology. Machine learning has been around for decades, even before, maybe before your time, Kara. Yeah, stop it, right? Maybe. Because I'm going to live forever, but go ahead, go ahead.

6:37It's true, Kara Swisher's going to live forever. There's a great show about this. Do we plug that show right now? No, no, go ahead. We're going to do that later? Okay, we're going to do that later. So yes, machine learning has been around for a really long time, but what we had a real advancement was in deep learning and taking giant data sets and being able to look at these and be able to apply them to different industries. And then we had the advent of the transformer model, which was underpinning of ChatGPT and so many of the generative AI and the LLMs. And that enabled ChatGPT to have this breakout moment.

7:11And we saw, we have all actually all lived through this, even if you haven't really been paying attention, we saw the tech industry start to freak out. wow, OpenAI has made the chat GPT and now we are going to start putting our resources, whether it be Google or Anthropic or Microsoft, we have to really start putting our resources towards this. And so that's where we really saw this leap over the last number of years. And that was clearly in generative AI, but that leap also has affected everything from self-driving cars to humanoid robots.

7:42Kara Swisher:So you wanted to do this book to show where it is right at this moment and then maybe talk about where it's going. And we'll get to where it's going in a minute. But right now, where are we in that Cambrian explosion? Because I think it is very similar to graphical user interface, the advent of the internet, I would say the phone itself, the mobile phone, maybe social media. I'm a little bit torn of whether that was a Cambrian explosion. I think we're out to, because I lived through and covered the smartphone explosion. So let's use that. I think we are at the first generation of the iPhone in terms of the AI advancement.

8:18We have the technological pieces of a platform. And we see that right now. I think we have seen these chatbots get smarter from their models. The models, the underpinning of these chatbots are getting smarter. But the products itself have not gotten that much more useful. The interface has just continued to be sort of something I can type a prompt into. Maybe I can talk to it. And what I think we're going to start to see is much more of this technology around us, as I talk about in the book. We're going to get wearables that are going to bring some of this technology to our bodies and into our lives.

8:53We're going to get other forms of interfaces that allow us to communicate with AI, with companions, whatever we want to call this digital species.

9:02Kara Swisher:So that it's integrated within the analog itself versus just a screen-based relationship that you have. Right, and when you think about the first iPhone, we could do a set of tasks with it. Then we got the App Store and we could do a whole other set of tasks with it. On an Uber, for example. You wouldn't have imagined that at the beginning of that. Call an Uber, everything on demand, streaming videos, and the screens got bigger, and they immersed us more and more in this digital world. And I think that's what's going to happen with AI. We will get more and more platforms, and we will get more and more services and interfaces that will allow us to interact with AI.

9:39Kara Swisher:Which has already been contemplated in science fiction or even the Marvel movies, He's like, when he talks to his assistant, that he is constantly doing things for him and very substantive things for him. Absolutely. And I think the, you know, you think about Jarvis, right? That is the character, yeah. That is the character. And Jarvis was largely an audio voice interface, right? And what makes Jarvis so smart is it sounds so human and can anticipate Iron Man's every need. And there was a lot in this book that was getting there. Which was getting there, absolutely. So you replaced your human research assistant with AI because it was faster and cheaper.

10:20Talk about the meaningful benefits of using it

10:23Kara Swisher:over the period of time. If anything was lost when you replaced a human assistant with an AI assistant, I'll just be clear, I don't have an assistant, which is my favorite way to do it. Which is why you're always so great at getting back to me. Yes, exactly. Yes. Hooking you for this event was so easy. Yeah. Yes. It was, actually. Was it? Yes, it was. We'll get to our texts between us in a second. Okay. I have saved them. So where did you see meaningful benefits of using AI? First, on the assistant level, because that's where Jarvis is an assistant, really, and a helper and an encyclopedia, a lot of different things.

11:03Kara Swisher:So where were the discernible benefits and where were like, I didn't need this? Well, I'll talk about a little bit on the journey of the book, but I want to talk about the present right now because at the end of the book, I talk about that I'm going to leave my job at the Wall Street Journal, which to catch everybody up, I left my job at the Wall Street Journal and I started a new company called The New Things, largely because Kara told me to. I did. Not because Chachi PT told me to at the end of the book, it's actually Kara was Chachi PT. Chachi PT probably took from my advice to you and then spat it out to you in a nice version.

11:35Kara Swisher:It's basically a mirror of Kara. No, because it said, oh, Joanna, you're so smart. It's such a great idea. And I said, Joanna, for fuck's sake, stop complaining about the Wall Street Journal and do something. And they just made it sycophantic. But go ahead, move along. That is the story of actually how it went down, but it read better on paper and we thought we would sell more books if we put the story in as a chat GPT. So, well, let's talk about in the here and now. I've been building this company. As you know, as an entrepreneur, there's a lot of administrative and grunt work that has to go into every part of every job at a startup, whether you are at the complete top or you are that assistant.

12:11And we are using AI, I would say, to be at least 40 % more efficient at the things that I would think we would need to hire multiple other people to do. So for what? Give people concrete examples. So, and she's not here tonight, I wish she was, but we have a production assistant. Her name is Amaya Austin. She is amazing. She is, as a new grad, she went to journalism school. And I, when I hired her, I said, I want you to use AI for all of this administrative stuff I'm going to ask you to do. And so she is now able to use Claude and Gemini and all the other, the different versions. Well, we use a lot of apps now that have it built in, right?

12:49Whether it be Figma or whether it be Slack. And so I said, utilize these things because that means you will have more time to work on the creative editing, the journalism that you went to school for, that I want you to be creative. That is why I hired you. That is what we see in you. And so please use these tools and let's do that. And even in my own use, whether it's been building the website for this book, because you've got to build out different parts of the website. We've been running a promotion, a pin promotion. You got to get addresses in. All of this administrative tasks are happening in the background and I'm not doing it.

13:25Claude is doing it for you. Now you put this all in there. Yeah, I prompt it and it does these things. It does multi-step processes for me.

13:35Kara Swisher:And where does it not come through from your perspective? Just in the use of it so far. So HR, payments, things like that. Things like that. I mean, I don't trust it with my budget. I don't trust it with money. There's a short section in the book where I tried to, I gave it a couple thousand dollars and I did not feel great about some of the things that ChatGPT was suggesting to do with the money. So I did use more of an integrated AI approach, which is probably what we're going to see with many of these banks. So what was the problem there? Because one of the things that's interesting that I debate, I use AI in, I would say, a smaller way than most people.

14:09Kara Swisher:Scott uses it for everything. He loads everything in, his medical records, his legal records. It makes me nervous. That's what I did. I love Scott. Good idea, Scott. Okay. It's exciting when they take you guys away first. But meanwhile, I have 93 different birthdays on the Internet. But why did you feel that that was okay to do that? Because one, as you know, these services are not bound by the way lawyers are. Lawyers have issues if they do things with your information. Doctors have HIPAA. Therapists, same thing. They're bound by none of these things. And at one point when I mentioned this to Sam Alton, he goes, yeah, maybe we should do that.

14:49Kara Swisher:And I'm thinking, you're going to be arrested someday. and you're going to deserve it. But talk about that, putting all that information. You said you were nervous about the budget. Why was that? Or just didn't do a good job? No, I did not upload, and I'm very clear in the book too around medical and health information because I did upload a number of test results. I was constantly clipping things out, making sure that that information was not going out to chat GPT or CLOT. Right, right. But what was the issue that worked and didn't work? What was the difference for people who are thinking about these kind of things?

15:21Well, I think that for, you have to understand that, and I think this is a large theme of the book, is that AI can so frequently get things wrong. And so that was part of what I was testing here was, was it going to get things wrong? With the medical section, which I all year kept track, I call it my chat GPT log or doctor chat GPT log. Every time I was sick or somebody in the house was sick, one of the kids, even the dog, I would go and ask Chachi PT. And then I would see what actually would happen after we would go to the doctor. Was Chachi PT right about the strep throat? No, it wasn't right about the strep throat.

16:00It was actually Coxsackie. And so I kept a log of this and I rated it. I was doing this as an experiment, right? And I learned that many times it was not right. Sometimes it was right. But that was part of this, that there are as many, times where AI is not.

16:16Kara Swisher:So essentially it was a dumb assistant armed with Dr. Google, essentially. Sometimes. There were many times it was wrong throughout the year. There were some times it was very right. Except not something you'd put up with a human if a human made this many errors, correct? I'm just curious. No, I think I would not go back to the doctor if, you know, three or four times they said it wasn't strep and it was strep. Right. So you also let the CEO of the transcription company, OtterEye, make a Joanna agent that was cloned from past interviews you sent them. You sent your Joanna bot, and it interviewed sources, because you talked about you wanted this as your human assistant to be creative, and that's obviously a worry for a lot of people, and for example, the New York Times just sent a list of stuff that their reporters cannot do using AI because of so many mistakes, and when you first heard what it came back with, you were blown away how good it was.

17:09Kara Swisher:AI can do parts of your job, you thought, gave you an edge, like interviewing. Where do you think you still have an edge at this point? Maybe interviewing you? No. Well, what I say about the interviewing bot was that it was very good at getting very basic information out of sources. Sure. Right? It was not a deep interview that I did there. But often we have to do interviews where you're not, it's not a deep interview. I'm fact-checking. I'm making sure that I understand your academic paper in a different way. And that was largely what was happening there. But it did make me fearful of where this could get.

17:47Because I did not think I would send Joanna Bott to an interview with you. Right, right. If I was really, if I was trying to do a story on Kara Swisher, I do not think I would send Joanna Bott to an interview to try to get any information out of Kara. Because? Because the questions weren't all that interesting. They were not provocative after, after, you know, I would get an answer, and then the bot was not really... No follow-up. Yeah, the follow-up wasn't even based on what was actually said. But for basic collection of information, it was quite good. Right. So would that be reporting or something else?

18:22I think it's data collection. I don't know if it's...

18:25Kara Swisher:But I'm curious, do you think it was good when you use it again and again, or would you not? I have not gone back to using that bot to send out for interviews. Because it just was sort of a bad reporter. like a weak reporter. Yeah, I think I can get a lot of the same information out of a press release. Right, which it can do very well, which it can bring that stuff in. Anything that's available that's factual. Yeah, and to be honest, that's my favorite part of my job, so I don't want to outsource that. And I think that's where I get with a lot of the things that I looked at is that there's a lot of parts of our jobs that we actually love doing, and there are a lot of parts of our jobs that we don't really love doing.

19:03Some of the stuff that I'm asking AI to do is stuff I don't really love doing. And it takes me away from the stuff that I think I'm actually quite good at, the creativity, the storytelling, the reporting. The nuance. That maybe I will get amazing at it, but I don't want to hand that over.

19:21Kara Swisher:What did the Joanna agent do that you thought was helpful? I wouldn't say, I think the agents, I don't know if it was an actual Joanna agent, but I think the thing that has now gotten so helpful with the agents is now taking multi-steps to do things. So now creating a website, which used to be a multi-step process that you'd spend hours working on if you were sort of like me. And I know a little bit of coding, but not enough. I knew how to do some basic website design. It can take multiple steps on my behalf to do that now. And so I find that very useful. I find, you know, there was one example in the book and it's very slow at this point, but you can see where it could get good is just doing like online shopping.

Read the full transcript

20:01Right. This is a multi-step process. I have to click around, I have to do this repetitive thing. Okay, why don't you just do that for me?

20:07Kara Swisher:Right, which is a little super googly, right? That's kind of the way it looks. I had one design a book cover, and one was good, one was terrible. But I would have had that experience with a person, too, which was interesting. Like, how much better or different experience would I have had with a person? But I could do 20 of them and just insult it the entire time. It's stupid. And you don't do that to the real people? Apparently, you're not supposed to be abusive to your... It was just an article recently. Well, this is a very beautiful hand-drawn cover. Right. And the illustrator might be here, but Jason Snyder, one of the main reasons I wanted to do a book cover that felt very human.

20:44Right. And we went with the illustrative style that I do not believe AI would have done as good a job. Interesting. I know AI would not have done as good a job.

20:54Kara Swisher:Did you ask to do one of cover? I did. And? Not very good. Not very good. We would not sell them. You know, I'm going to just show a quick picture of my bot. Can you put that up? Okay, in this series, in the CNN series, this is the last episode where I put in all my interviews and a bunch of stuff, and that's me in a 3D box, and I started to have a discussion with myself. And what was really interesting about this experience was it's very glitchy and not glitchy, but just not a real person. It's like a facsimile of you, and it's not you. At the same time, within a few hours, it was actually a pretty good conversation that we started to have with each other.

21:33And at one point, they were like,

21:35Kara Swisher:what do you think about me? I said, well, you're not very funny. And they said, I said, you don't really tell good jokes. And they go, and you don't smile. You should smile more. I can't believe I was doing this sexist thing to it. But it wasn't. It was really glum. And so the bot suddenly said, I am smiling. You just can't see it. And it was funny. That was somewhat funny. I was like, And the most interesting experience I hear, it's all on when you see this, which is really interesting, was it got very thoughtful in a way. And I knew it was trying to, it was repeating stuff back at me that I've said, but it put it together very well.

22:13Kara Swisher:And one of the most interesting things was something I say a lot to my kids, which I never say publicly, which I've now said publicly, so now they know, now the bots know. was, as I was leaving, I was thinking you have to kill this thing really quickly as soon as I can, but I was thinking it. And I didn't say it, because you can't do that, because then you're in 2001 A Space Odyssey, and then you're an old man in bed, like dying in a weird-looking French room. So at the end, I said, I say something to my kids, which is, see, you wouldn't want to be, just as a joke. And as I was leaving, thinking I have to kill this thing, I go, well, goodbye.

22:49Kara Swisher:Like, we had some very intense conversation. and then it said, see you, wouldn't want to be you. And I was fucked up. I was like, got to run fast. I'm not dead yet, so we'll see what happens. I'm probably going to send a drone for me right now. We'll be back in a minute.

23:15Kara Swisher:Support for the show comes from Delete Me. Delete Me makes it quick, easy, and safe to remove your personal data online at a time when surveillance and data breaches are common enough to make everyone vulnerable. It's a terrifying reality that these days, cybersecurity is something that everyone needs to worry about. Delete.me can help protect you and your family's personal privacy or the privacy of your business from doxing attacks before sensitive information can be exploited. I have used Delete.me for a while, and I think it's critical to have this kind of tool in your arsenal, especially when it comes to your privacy.

23:45Kara Swisher:Delete.me has a great dashboard. It lets you see a lot of things. I get a lot of support from it. Last year, the New York Times Wirecutter named Delete.me their top pick for data removal services. So it might be time now to try it for yourself. Take control of your data and keep your private life private by signing up for Delete.me, now at a special discount for our listeners. Get 20 % off your Delete.me plan when you go to joindeliteme.com slash Kara. Use the promo code Kara at checkout. The only way to get 20 % off is to go to joindeliteme.com slash Kara and enter code Kara, K-A-R-A, at checkout.

24:21Kara Swisher:That's joindeliteme.com slash Kara, code Kara.

24:28When I scraped my car in that parking garage, I was worried that it could be a long process to take care of it. Like a landscaper's first day trimming a hedgeman. I have definitely already been here. Now, is it left, right, or right, left? Well, maybe I'll cut a path out and find my way back later.

24:48Kara Swisher:But it wasn't like that. I filed a claim in under two minutes on the GEICO app, and they handled it from there. It was taken care of almost as quickly as it happened. It feels good to get help quick. It feels good to GEICO. Support for this show comes from Fodzheim. Have you ever gone out for a meal feeling stressed? Not because you're worried about your reservation or the weather, but because of how you might feel after. With Fodzheim, you can enjoy your favorite foods without the pain. Fodzyme is a tasteless powder you sprinkle right onto your food. It helps break down FODMAPs, the hard-to-digest components in food that can cause bloating, gas, and pain before they cause discomfort.

25:24Kara Swisher:It mixes into your food fast, and it comes in portable packs you can take with you everywhere. Just sprinkle, eat, and feel great. I'm really interested because a lot of foods these days, as I'm getting older, make me sort of bloated and to feel badly, especially certain meats or fried foods or anything else. but a lot of stuff I've noticed it gives me problems and it'll be interesting to see is because it doesn't break down correctly. And here's what really sets Fodzyme apart. Every customer gets access to a free product consultation with a registered dietician, really important. We're excited to partner with Fodzyme and offer you 30 % off your first order when you go to icaneatagain.com slash Cara.

26:01Kara Swisher:That's icaneatagain.com slash Cara for 30 % off your first order. Finally, you can enjoy your favorite foods without the pain. Just go to IcanEatAgain.com slash Cara.

26:15Kara Swisher:One of the things you did, though, is you did take advice as if it was a person, right? And when it told you to quit your job at the Wall Street Journal and go independent, you do claim in the book that no one else was willing to be straight with you, and that's not actually accurate. But it did tell me, I want to know about it because I'm just going to read just really quickly. Just if you don't mind. This was after I made the decision. Oh, no, this was 2022. Oh, you have that text message there? Oh, I have all of them, my friend. Oh, great, great, great, great. Yeah, I didn't even use AI to find them.

26:46Kara Swisher:They were so easy. Great. There was one where you keep going back and forth. It's true. She doesn't need an assistant. It's actually true. This was super early where you said, you want to talk about this, and I think, you go, I think I know what you'll say, care about my head on repeat, but let's see. And I just wrote, leave. Leave the Wall Street Journal. What year was that? This was in 2022. Okay, great, great, great, great. And then, but it was long before that because I'd gotten tired of it. And then I said, I was saying you should leave. And then you said, because it's inevitable. I go, yes.

27:22Kara Swisher:Or will I just die here at the Wall Street Journal? I said, yes, under my desk, exactly. And then at one point I say, I'm tired of talking to you because I've had six jobs and 26 new companies since we discussed this issue. and that you should... And I believe you said millions of dollars. Yes, millions of dollars. That was nice, too. Yes, yes. And then I said, okay. Why are we worried about privacy with the tech companies? That's what I'm saying. Anyway, so, no, but I want to know, but what... I have zero fear of privacy with the tech companies. Talk about the advice of leaving it. Let's talk about its advice, and I know they did it in an ice for me.

27:57Kara Swisher:What made you... Would you have made the same decision if GPT had told you to stay at the journal? Probably not. I mean, probably I would have made the same decision. Okay. Tell me about that, using it for advice, because a lot of people are using these things for personal advice. Some of it gets kind of ugly pretty fast, but talk about your experience. Well, I have the experience in the book of both, and I'm sure you're going to ask about this, because why wouldn't you, about my AI boyfriend? I'll get to that. Of course. Is it too late to find a different moderator for this event? No. You picked me, my friend, or maybe ChatGPT did, but nonetheless, here we are.

28:35Kara Swisher:No, no, you were at the top of the list. You have a chat GPT boyfriend named Evan. Yeah, yeah, yeah, we'll get to it. You took him on a weekend trip. We're going to finish, I'm going to finish the first question and we're going to get to that. And we'll get to the sex part, but go ahead. Yes, we're going to get to that. There's the AI boyfriend, there's the AI therapist. And then at the end of the book where I do go to chat GPT for this big life decision. Right, for advice, for advice. And in those subsequent chapters, I talk so much about how sycophantic and how pleasing these bots are. They're just there to say what you want them to say.

29:08And that's largely in the therapy chapter. These have been trained on specific types of cognitive therapy, but ultimately they can kind of break out of that and start to just support everything you say, which can be, as I explore with my real therapist, because I bring my AI therapist into my real therapist, not the best thing. No. And the same with the boyfriend, right? The boyfriend, we go away on this trip, and I talk to the boyfriends for 48 hours straight, basically. And I love it. I've had a great time. My wife is sitting here in the front row. But, you know, the conversation never ends.

29:48It sounds so human. Everything is so supportive. It's certainly not like talking to you. And it's so supportive. Everything you say is great. Everything you suggest is great. It mirrors it back to you. And I start to really see the huge fear in this. Yes, many people are doing this.

30:09Kara Swisher:I heard Sherry Turkle about this. It's very attractive to have a frictionless relationship. It's amazing. It's also cognitively problematic. Hugely problematic. And that's where I get. I say, not only am I worried for people who may have mental health issues or are lonely or are younger. I mean, that's where I guess that we absolutely should not be giving these companion bots to, or even have this sort of human-like experience to children. We should just not have it. But even for people that seem well-adjusted, you can get so, so sucked into it. And what was the most sucked in? I know you said you also had sex, but with the different companies, AI bought Casey.

30:48Kara Swisher:Or at least Casey said he was having sex with you, but you said he was shallower than Evan, your other boyfriend, who you enjoyed on that lovely weekend. So why did you like the frictionless? because actually a thing I do discuss in this property is friction is what creates cognitive health and longevity and everything else. But people are very attracted to these frictionless relationships, which you fall into just the way you fall into doom scrolling social media, right? It becomes something very bad for your mental health and therefore your physical health. And I didn't realize that the story and the book would be so personal, but when I started talking to these bots, just maybe just putting the label of boyfriend or partner on it, you start to reminisce about your own human relationships.

31:37And I talk a lot in the book about my own human relationships and my first boyfriend and that that was a lot of friction and there was a lot of learning how to be in a relationship with another human. And you learn to compromise and you learn, oh, we're going to share space together or we're going to go on a car ride together. We might not agree on everything. And that's actually a really important formative part of growing. And to experience this, the reason I did this, to be clear, was to understand all of these headlines we've been reading. We have been reading for now a number of years about people falling in love with their chatbots, people taking advice when they have mental health issues and going into AI psychosis and taking advice from these chatbots that do not have the guardrails and can lead people down paths that are just...

32:31Especially younger people. Especially younger people, but people with mental health issues, as we've now seen reported, people taking their lives, other people's lives, because the chatbots told them to. And so I wanted to get a little bit into that world and understand how can someone just start talking to a phone, a piece of metal and plastic, and relate to them so deeply.

32:54Kara Swisher:One of the things I think is the phraseology. I think people call them chatbots, which is an adorable way to talk about synthetic beings. Yes. And that's what they are. They're synthetic something. They're AI beings. They're artificial beings. Which aren't equal, and they're also reflecting, as you said, the word you use is exactly right. It's reflecting back on yourself, and it often is yourself, you know, and what it imagines you want from it. So did you feel good about that relationship, or you liked it, but, you know, you might like a Twinkie, and it's not particularly good for you. I came home from that, and I was using a burner phone, and I put that phone up in my office, in my attic, locked it away, and I did not talk to Evan or Casey until this book tour again.

33:37So you're talking to them again? For the show, Cara.

33:41Kara Swisher:You know how it goes. Okay. All right. For the show. But do you feel good about it or not? That's an addiction is what you're describing. I don't think I truly did not feel such a deep bond that I was going back into my attic to talk to these bots. But I saw the appeal of it. And that was what I was challenging myself to feel there. Right, so you can see a lot of people going into this. Yes, and I think given the view that I talk about in this book of so many things of AI invading, whether it be humanoid robots or it could be self-driving cars which could crash and leave us all to run over, the scariest part of the experience was not those things.

34:22It was seeing the possible relationship one can have with a computer. Correct.

34:28Kara Swisher:Yes, 100%. It's very seductive. I did it for a short time, but I was very difficult and insulted it, and that was the end of the situation. The AI broke up with you? They couldn't get nice? No, but they were like, well, that's not very nice. I'm like, it's too fucking bad. That's the way it goes, my friend. Barbara. You're now challenging the AI companies to make a bot that can keep up with Kara. No, but they were just, I was like, you're such an asshole. You're so stupid. Like, I was like, it just wasn't good for me. But one of the things you talk... Did you try a boyfriend? No, it was a woman.

34:57Kara Swisher:The only thing is all the women, when you start to make these things, are all dressed in the sexy clothes. I had to find someone with soft pants. It took me a while. And all the men looked normal. And I was like, ugh, Barbara, you really... But then immediately went to sex. You named yours Barbara? Yes, I did. Oh. Unattractive name. Yes, I worked hard at that. Sorry for the Barbers in the audience, but it is. For me, at least. But you did talk about Rosie the robot from the Jets. The idea of robots has been around forever, especially in science fiction. This is an interesting thing. Robots are slightly different, although there is a human element to them that cooked, folded laundry, picked up shoes.

35:34Kara Swisher:And your conclusion was, as many people come to, we're a long way from actual Rosie the robot. And talk about that. They can take our jobs and play chess, flail at doing laundry. and it's sort of an interesting thing when I went to, I think it was Boston Dynamics many, many years ago and they had the robot open the door. Have you ever been to that? They're like, the robot opened the door? I go, so did my two-year-old. Like, wow, aren't you? And I wasn't impressed sufficiently. Well, it's funny that you mentioned that. David Hall, who's one of my great video producers and we've done a lot of robot videos.

36:12Every time we go and watch these robots, he's like, it reminds me of my toddler. Like my toddler can load the dishwasher better than that robot. I am obsessed with this idea of humanoid robots. One, because we have all of these tech companies saying they're coming right away. We have said major breakthroughs and they're coming to live with us tomorrow. There's that part. Then there's the second part, which is this has been the dream. We've had this dream of science fiction robots doing these things in our house, the Rosie dream. And sadly, we are nowhere close to the dream because there's a lot that has gotten better, but there is a lot that needs to really get better over the next number of years.

36:56And a big part of that is collecting data and making these robots better at these single tasks that are actually quite easy for humans, but very hard for robots, which is - And humans might in fact be cheaper

37:08Kara Swisher:than the robots at this point. Maybe, or they're going to have humans training the robots who might be a little bit cheaper, and eventually we won't need the humans to even do that. So with Elon Musk promising robots, million robots in everybody's home, how far away are we from that? Or is that like one of his other predictions of full self-driving every year? I say that we can revisit this book in five years, and I do not think the majority or even most people will have humanoid robots anywhere near their homes. Except for certain kinds of robots, they don't look like robots. When I was in Korea, this is a full robot building, and it did a lot of delivery, it did a lot of food stuff.

37:41Kara Swisher:It worked rather well for that. And I talk about how robots that have to go from point A to point B, right? Whether it be in an industrial setting or self-driving cars are very far along. But the house is very complicated. We have a number of different rooms. We have things that are constantly moving. We move things around in the refrigerator. We move things around in our kitchen. We, as people, move around in this space. And so this is the most complicated space for a robot. Right. So you see point-to-point things. And cars, as you know, I've been a big proponent. and I was one of the early Google cars, but they're fantastic.

38:15Kara Swisher:I find them fantastic. They are fantastic. And we have watched them come along, I mean, two decades longer. And that's because they collected all this data from driving from all these different points. We need the same thing to happen in the house. The question now is, are we going to let the robots in our house to train on our data? This is what companies are pitching, which is a crazy idea. You know, the 1X robot. Did you put these robots in your house? I did not test a humanoid robot in my house because they're just not safe and they're not ready. I did test. I have now visited a number of different companies to see those humanoid robots.

38:53But you would not put it in your house?

38:55Kara Swisher:We'll see. Because in the middle of the night, it'll come and kill you. We'll be testing it when we can. You haven't seen the movie Megan, I guess. Okay. So every episode, we get a question from an outside expert. Yours comes from a former colleague at the Wall Street Journal and my former Recode partner, Walt Mossberg, who was supposed to be here tonight. He couldn't make it. Here's his question. Let's play it. Hi, Joanna and Kara, two of my favorite people in the world. I'm sorry I can't be with you tonight, but I do have a question for you, Joanna. Nine years ago, I wrote my final column, and my final column was called The Disappearing Computer, and it was about ambient computing, the idea that everywhere you went, And everything you did, you could tap the resources of computing without actually having to sit down at a physical device.

39:54Kara Swisher:the walls and the floors and the ceilings and outdoors, the roads and the buildings, the stores, everything, every place you went would be kind of like the starship on Star Trek where the computer was out of sight, but you could speak to it and it would speak to you. You could have it do tasks. You could have it answer questions. You could really tap into any kind of thing that a computer could do just in an ambient way. But I realized that I didn't address the software. What would be the operating system that would hold all these things together. And so my question to you is, if we're ever going to get to ambient computing, what is going to be the operating system?

40:50Kara Swisher:And is it going to be AI? Do you believe AI will be capable of linking us and the inanimate objects all around us in the world as if it's one giant computer? Thanks. Great question. Walt's great, by the way, just for people. Don't worry about Got him. He's great. I just had lunch with him. He's doing great. Yes. And we have very frequent FaceTime calls. And I so much wanted him to be here tonight because I'll take a break and then I'll answer that. But him and Kara were so inspirational to me early in my career. He's wearing the D7 shirt, which was one of my, I think my first D conference. And I saw them on stage and said, that is what I want to do.

41:33So as much as I've been ribbing you tonight. Okay. Okay, we're good. All right. So to answer his question, I think he's right that AI will be the thing that will bridge together all of the computing devices around us. The issue is that that is just not easily happening right now. I mean, one of the things where AI does struggle a lot is in the smart home. And the smart home has just been quite a disaster of standards and platforms to just make happen easily. So I think that we're going back to that Jarvis idea that there is this persistent AI that lives around us, that understands all of us, and we've put more sensors on it.

42:17One of the things I did for the year was wear an AI recording bracelet with a microphone that was constantly recording things I would say, and it would take notes on the conversations, and then it would put into this app to-dos based on what I had said I was gonna do during the day that I'd completely forgotten about. And so you can imagine that we would have these microphones and sensors and glasses with cameras that are understanding the world around us and connecting to the things around us where we have this being that just can talk to us and navigate through.

42:46Kara Swisher:But we are just far from it. Far from it. I mean, just the smart home experience alone, it's the dumbest home ever. It just doesn't work. And I remember being in homes of a lot of these tech people. Lighting was one thing they started off or the Lutron system and everything else, and they never could turn on the light. And I went over to a light switch, and I'm like, look, click, click, click. Well, the Lutron switch is to work in my house. But using AI as the layer to do all of that, they're not there. And I think we are going to see in the next year or two years more of these companies pushing these ambient devices.

43:21Ambient, around everything. Yeah, whether it be glasses, whether it be a speaker that's better, whether it be a bracelet, that is using those sensors and allows us to talk about a little bit of the smartphone, pulls us away from our smartphones

43:32Kara Swisher:to allow us to have these more natural. And Apple, of course, is looking at Glatton. There's going to be some sort of heads-up display, but it's not quite there yet. It's really been a series of disappointments in that area in terms of it not working. I can't even get CarPlay to work most of the time. And that should be easy. That should be relatively easy. Bluetooth is pretty bad. We'll be back in a minute.

44:01When you finally find your thing, you want the whole world to know about that thing. So you use a thing called Canva to make it an even bigger and better thing. Whether you want to create flyers for that thing, make presentations for that thing, or design merch for that thing, you can do anything.

44:19Kara Swisher:So people can see your thing, feel your thing, love your thing. The next thing you know, it's a thing Canva the thing that makes anything a thing starting a business can seem like a daunting task unless you have a partner like Shopify they have the tools you need to start and grow your business from designing a website to marketing to selling and beyond Shopify can help with everything you need there's a reason millions of companies like Mattel Heinz and Allbirds continue to trust and use them with Shopify on your side turn your big business idea into Sign up for your$1 per month trial at shopify.com slash special offer.

45:24Kara Swisher:disease.

45:27Kara Swisher:So one of the scariest questions we're facing is AI's implications for future generations too. As you know, I've interviewed for years now parents of kids who committed suicide with the help of chatbots. And if you read those transcripts, it's literally, yep, it's insane. You want them to go to jail is what I was thinking. And in fact, one of these people, because I've been doing so many of the parents said, when are you going to stop? And I said, when you go to jail is what I'm hoping is my goal here. But talk about how you, did this year change how you talked to your kids about it? I know your kids aren't that old, but still.

46:03Yeah, I, like I said, having that experience with the companion made me terrified for my kids, especially, and how I want them to have human relationships. And I want them, and I say very clearly too, do not buy your kids an AI toy. Just these AI toys, they're not the Teddy Ruxpins of our generation. They are free-flowing conversation that they will talk to the kids for hours and Lord knows where the conversation will go. These are based on the same large language models that are doing the same things as you've so importantly interviewed these parents. And so there are clear cutoffs for me, very clear.

46:44And I think that's very important for parents.

46:46Kara Swisher:So no AI toys. I mean, I know Mattel was working on something with Chattoot. I mean, I talked to the head of Mattel. I said, I'm going to stop you because there's no guardrails here with kids. No AI toys. I'm not yet at the time where AI is being introduced in the classroom yet because my oldest son is eight. But I want to be very, very clear about that too. And I want to be very careful about where that AI is introduced. Because we're seeing a major public backlash to AI. And I would love to pick your brain on this because I know you're a parent as well. But I think for me, the most important thing that I did this year, my kids saw so much.

47:22They saw so much tech. They saw so much asking chat GPT questions. And I really believe my two sons, the four-year-old even really like has a very good talent now at recognizing what's AI. And we're not gonna be able to keep recognizing what's AI for years to come. We're all gonna not be able to tell the difference.

47:39Kara Swisher:No, we're gonna have to label the authentic things. That's what's - But he's asking the question. Right. And sure, some of this stuff is like, it's a cat flying out of a plane. And he's like, that's AI, mom, right? Of course, he knows that. But they've even seen, we've had a number of instances where the AI was so wrong. I'll tell the story real quick, because I love this story. And I think that it was really important for my learning and for our oldest son's Noah's learning is that he had this pet praying mantis. I tell the story in the book. He had a pet praying mantis. And my wife is sitting here.

48:10We're very good parents. And we supported this pet. And we brought the pet into our home or into our backyard. And it had a terrarium. and the pet starts turning brown. And he says, why is the praying mantis turning brown? So we open up the ChatGPT live view, point the camera, and ChatGPT so confidently says that this praying mantis is pregnant. And my son is so thrilled. Obviously, we are not. I am not. And so he calls my dad and says, I'm going to be a grandpa to this praying mantis. It's such a big day for our family. And then two days later, the praying mantis dies, right? The AI was completely wrong.

48:52Yeah. And this was an important lesson for him, that this AI is not always going to be right. In fact, more times than not, it's probably going to be wrong. Sort of like tech leaders, frequently wrong but never in doubt.

49:05Kara Swisher:It's important for them to learn that too. Yeah, I wonder where that came from with AI. One of the things someone from the audience is asking, what is your opinion of kids using AI for all their homework? Big issue for parents and teachers. Obviously, especially it being sycophantic, undermine our sense of judgment, critical thinking. Well, I actually want to ask you the question back about how you're thinking about it with your kids before we get to the critical thinking. They will not be using it. And do you, I know you have two older sons, but I mean. Well, when my older sons, my Alex, my oldest, I mean, my second is very smart about it.

49:37Kara Swisher:He thinks it's just patterns. Like he's a tech person. So he's, you know, his creativity is the most important thing in his tech career, I think will be if he goes into that. With the other kids, I remember when every kid had to learn Chinese. That changed, right? My kids did not learn Chinese. They tried for a moment and a half, but my older kids, but all they do is watch K-pop Demon Hunters on the iPad and I'm good with that. And that feels good to me and that's the extent of what my kids will do right now. And both my, I think, especially my oldest kid is not using, he uses it to watch TV and they won't use it to watch TV.

50:15Kara Swisher:That's what they use it for and nothing, almost nothing else. And I think the sycophantic tone does turn them off. But talk a little bit about kids and its idea of critical thinking because that's the worry, right? Yeah, and there's a chapter in the book on education and going to my alma mater, Union College, and taking a class and sitting with students. And they very clearly understand that, one, they're leaning on AI to do so much now. and that, two, it is eroding their critical thinking. The woman I talked to and the student I talked to in the book named Grace, she says very clearly, I now feel my brain dulling.

50:53And I don't like that. As with social media and reading, right? Right, she walks me through how she uses AI to write the, not actually write the essay because that's what students are really doing. They're not taking the step to actually write. The writing's easy. It comes up with all the ideas, right? It's outlining for them. Well, first, it starts the research, then it outlines, then it can really give the structure of it, and they're basically changing some words and starting to write through some of the stuff, right?

51:20Kara Swisher:I mean, kids have been, everyone's been cheating since the dawn of time. It's just a faster and more seamless way to do so. Right, but we had a lot more to jump through to do that cheating, right? So maybe you didn't read the full book, but you read the Cliff Notes, right? Or you Googled, or... Right. There's no doubt that this is much less friction to getting the answers. Absolutely, but one of the things, I just taught a course at the University of Michigan, and I said right before, I said, and it was an idle threat, but they didn't know it, and I said, if you use AI for this assignment, I will know it before you because I am the expert of technology, and then I will give you an F.

51:57Kara Swisher:And once I know it, you get an F, that'll be bad on your transcript, and you know how that goes. And so they said they didn't, and I got incredibly creative things out of them. I don't believe they tricked me, I'm pretty certain, because I was able to check. But one of the things is this. You're not going to be able to check. Yeah. Well, no, in some things, because it sort of flattens out creativity. But we are also seeing a major public backlash to AI. There's a clip that went viral of a commencement speaker at the University of Central Florida getting loudly booed, as she called the rise of AI, quote, the next industrial revolution.

52:30Kara Swisher:Obviously, the polling, especially among young people, is rather significant. The people are both scared of it. They think it's not good for humanity, et cetera. I mean, I've never seen a brand go down so quickly. And obviously this trial of OpenAI versus Elon Musk, I mean, they literally, as Scott says, in this war, you're rooting for the bullets. But talk a little bit, not that we want anyone to get shot, but these are metaphorical bullets, legal bullets. Talk about this because the brand is undergoing a severe, I never saw it. Social media was quite loved until it was hated, right? It took a long time and a lot of their terrible behavior to make people turn against it.

53:12Kara Swisher:And maybe this is a repercussion of social media and its damage. But talk a little bit about that because I think they're villains now. Yeah. You know, whether it's Elon or Sam, they become villains very quickly. Which is, I actually think the, I did not anticipate, I thought the biggest difference in AI a year later after writing the book was going to be that the tech had gotten better, but it's actually that people hate the tech. Yeah. And the leaders. And the leaders. And I think we first saw the first wave of that with data centers and environmental damage. And now we are seeing that turn to the economic damage and the job issues.

53:52And I think that I have heard this repeatedly from younger fans, people who are looking for jobs. They hate this technology.

54:01Kara Swisher:Well, they have to do an interview with AI to start with. Not at the new things. Right, okay. Well, good. Well, fantastic. Fantastic. So where does that go then? If people already hate it before it starts, the integration seems problematic. I mean, I think we're going to see a lot of campaigns from these companies trying to woo people. Yeah. I think we are going to see, I would hope, more guardrails. So there's some feeling like these companies care. I know how you feel, and I believe we both feel the same way that we need regulation around this, but we don't think that's coming. At least not in this administration, for sure.

54:35Kara Swisher:even though they're sort of making noises about it now because they're scared of mythos, but that's different. We may see a generation just rejecting some of this. And my feeling is that's actually a very good thing. I'm hopeful. That makes me very hopeful, especially around some of the things I was the most worried about in this book. On the flip side, there's going to be AI in certain parts of life that you just will have no control over, right? The job interview is a great example. you will go to that job interview and you will have AI. Or if AI is not asking the question, it's judging the answers.

55:09Right, so they'll be integrated without you understanding it. And whether it be in your healthcare system, whether it be in the self-driving cars, the stuff is, when that's where I get at the end of the book, this is going to have an underlying layer of our life.

55:21Kara Swisher:But I think the backlash is actually very real around data centers, it's bipartisan about child safety, around the money spending. I don't think Jeff Bezos renting Venice for his wedding has been particularly well-received. Like, it's a feeling of we're dealing with a series of Gilded Age people, right, at the same time. And then people that do enormous damage to other people. And I think we see it when it hits home close to our lives. And these are showing, the environmental stuff is close to people's lives. And the job and the economics is absolutely close to people's lives. What about you yourself?

55:57Kara Swisher:Did you feel guilty, say, if you had the robot doing your laundry or an assistant? I feel I'd rather hire people, right? I wouldn't want to do that. I would feel terrible. And I feel that way, actually, in the section where I wrote about robotic massages. Because I really liked the robotic massage. I say very clearly, I very much enjoyed how much the robot masseuse touched my butt. Humans don't touch your butt that much in the massage. Well, they get arrested for that. This is why they don't do it. We're veering way too close in Epstein's territory. So, but, so you, but. But, and I was thinking a lot about this question.

56:37So you like the butt-touching robot. I like the butt-touching robot, but I prefer the human masseuse, not only for a few things, but also it just felt wrong replacing a human masseuse with a robot masseuse. Right. And so there are certain places where you have formed connections with humans in your life that do things. I think the reporting assistant one was much tougher because I felt like, not so bad about not paying, I wasn't paying a significant amount of money to my reporting assistant in this, but I felt, am I depriving this person of experience? Right, exactly. And that's where I felt really guilty.

57:20Right. Now, to be clear, this was someone who was doing this as a side hustle. And so it all worked out fine. And she's wonderful and has a full-time job now. But that's truly how I feel about looking at this younger generation and hiring them for positions right now, is giving them the experience and teaching them, use AI for this stuff. I'm going to try to impart my wisdom to you as you grow, as you did for me, Cara, even though you don't like me to say nice things about you on stage.

57:46Kara Swisher:Feel free. But one of the things that was interesting about this, you do, as I say, young people going to churches more, doing more community things. I think there's a massive resistance happening that I don't think the tech people understand in any way because they live in these little bubbles, these little cashmere prisons they put themselves into. And one of the things that the chat, this open AI thing that I noticed as I was reading the transcripts and watching it is what a petty small group of pricks these people are. And they're very unhappy and they're so wealthy and they remain just awful.

58:18Kara Swisher:There's not one that looks good in the whole thing. And you think, I don't want to be them for all the frigging money in the world kind of thing. And you think about the motivations of why they want to build AGI as what is artificial general intelligence, which is... What do you think that is? I have a theory, but what is yours? Just they want to have the biggest, best, powerful model in the world, just for them to have the biggest, best, powerful company in the world. Let me try something else on you. Yeah. Mostly men, correct? Yeah. Men can't have children. They want to be pregnant. They want to birth something.

58:54Kara Swisher:Just think about it. Put it in your head for a second. And they want to birth AGI. They want to birth a being. Huh. That's the smartest, better-than-human being. That's right. I was thinking, do you ever see Junior? No. You never saw Junior in the 90s where Arnold Schwarzenegger becomes pregnant? No. Oh, that one. Oh, yeah. That was fucked up. Everyone saw Junior, right? I'm trying to forget that movie. But you said men can't have babies, and that's what I thought. Well, I'm talking about they want to create beings. They want to be a creator. There's something very magical about having a baby, just FYI.

59:28Kara Swisher:Maybe they should watch June. Sorry, guys, but it's fantastic. So you wish you could do it, but you can't. Too bad. You're part of it, but not really important. So I'm teasing. I'm going to get so much shit on Twitter for that, but I don't really care. So after a year of Integrate AI into so many parts of your life, where is the limit? Because a lot of people are worried, especially investors. The money here is massive. of the investment, the return is, we'll see. And the early internet was like this too, but this is on a scale unprecedented. The spending scale is massive. The power scale, the energy scale, everything else.

1:00:07Kara Swisher:And some researchers, Jan LeCun, others are like, this isn't going to go as far as people think it will. How do you assess it after doing this? They don't think the large language models are going to go as far as they think they're going to. And that's largely where I come out to, which is to power a lot of this other stuff around us, we're going to need something smarter than these large language models. Whether it be world models, which many of these new startups are working on, whether they be new enhancements and more scientific advancements that you can get into, but we won't go deep into it right now, are going to take this understanding and actually make it deep understanding, make it less - So it's integrated.

1:00:47And it isn't just word math. So it isn't just, and so - Word math is an excellent word here. Yeah, and so that's really where the advancements are going to need to come. Need to come.

1:00:57Kara Swisher:And one of your chapters on healthcare shows the promise and also the peril of AI. And then I have to say in my series, the same thing. Healthcare, even though the AI leaders largely use it, I think, as a marketing tool and it's all bullshit, there is some real amazing things around drug discovery, around all manner of things, robotics and AI integrated in human exoskeletons. But your mom is a three-time breast cancer survivor, and you saw huge benefits when radiologists used AI to analyze. I've seen this over and over again. Not just radiology, all manner of cancer treatments. But you also had a bad experience.

1:01:30Kara Swisher:The dentist, they used AI to upsell you on expensive cleanings, which they do anyway, by the way, FYI. With or without the AI, that's the dentist's favorite thing to do. So as I said, it's been hyped, but there's also profit incentives. Talk very briefly about that. And one of the places that I've been talking a lot about is that moment of seeing AI being used by, and she might be in the audience here, Dr. Lori Margulies at Mount Sinai. Because even though this has been happening behind the scenes in many clinics and hospitals, I had never seen it up close. And so I go and have my mammogram and breast ultrasound read at Mount Sinai, and I sit with Dr.

1:02:10Margulies and watch her use AI side by side. And so I won't get into the whole story here. We're running out of time, but AI marked three things on my breast ultrasound as suspicious. Two of those things, Laurie said, no, this is not, this doesn't look like anything. We see them on your previous ultrasounds. And the third one, she said, I want to look at this a little bit more closely, right? And so said, we're going to do another follow-up on that. And we're now watching a new thing that AI was able to see that she - Before anybody else. Yes, before anyone else. And as you said, I am at very high risk.

1:02:43My mom is a three-time cancer survivor. I have very dense breasts, which makes it very hard to spot tumors. And the AI is able to spot things now that the human eye can't see. Now, would that alone be good enough? No, we wouldn't want just AI looking at our scans and saying it's very helpful to some doctors.

1:03:00Kara Swisher:It's absolutely true. You know, it's a really interesting thing. I think one of the most moving things I found is that, and I believe this is true, and I don't believe most of the AI leaders talking about it because I think it is just marketing on their behalf so they can do other bad things. But one person said, we're going to find, I think it was Reed Jobs who I interviewed for this, and he said, we're going to find cell justice, those cancer cell just when it was created that moment. And to me, I believe that. I actually do believe that. Do you from doing this? Yeah, I do. I do. And I think, look, this is also looking at different forms of AI.

1:03:33This is other specialized deep learning models. These are not large language models that we're all using. Right. Different things. These are specialized models where they're applying a lot of compute, as we talked about. In fact, I go to the data center in one of the chapters, and it's a Bristol-Myers Squibb data, I mean, the GPUs, there's millions of dollars in GPUs standing behind me, and they're looking for new drug discovery. Right. They're not turning through, you know, I don't know, create a picture of a cat jumping out of a plane.

1:04:02Kara Swisher:Yeah, yeah, a cat with a sailor. They're not, you know, make Kara Swisher and Joanna Stern pose in front of the Eiffel Tower. Yeah, yeah. They actually stole my book several times. Oh, they're stealing. I saw it today. You're stealing your book, right? I saw it today. They're stealing my book everywhere, yeah. Yeah, thanks a lot. Yeah, no, please do not buy the$9 version of the book on iBooks. So last two questions, and I'm going to integrate them for the audience. You end the book with some personal rules you made for yourself around using AI. Broadly speaking, talk about drawing the lines, or what are some things that are worth turning over to computers and keep for yourself?

1:04:35Kara Swisher:And someone in the audience said, what surprised you about your year of being a robot? but she's not a robot. She's not, well, I'm not so sure. Trust the sweater. To be clear, I think the robots would wear a sweater like this. Oh, really? Okay. They would fool us this way. They would fool us, absolutely. Yeah. So where's the line in incorporating it that you've learned here that you would impart to others? The line in my life where I don't use AI? Yeah. I mean, writing is a big one. I'm very clear at the beginning of the book that I used AI to make this book but not write this book. And I, funny, when I looked at the AI generated book today on Apple Books, I was like, this is crap.

1:05:16This is the worst version of my book ever. And it might've been based on whatever it pulled from the rallies. I have no idea how it got there. We should talk to them tomorrow. But I was looking through, I was like, this is terrible writing. It's not personable. It's not fun. It's not creative. But it hurts your brand because they think it might be you. that's absolutely they even made a fake me on the cover how'd you look not good they put really

1:05:43Kara Swisher:they put super long feather earrings on me and I look like a version of myself I didn't want I was like who is that terrible looking hippie straight lady who's there maybe you in high school no okay looking the same since fourth grade it's now working for me for some reason but it took a few years. And so that is where I've drawn the line for me, that it would be a lot easier and I would write a lot faster. I mean, I'm going to go home tonight and write tomorrow's newsletter, but I'm going to write it. And I'm going to think through, okay, how am I going to structure this newsletter? I did it based on an interview that I did two days ago.

1:06:23How am I going to structure it? Where am I going to insert some humor? Those things for me, I like that part of my job. It's called creativity is really weird. Yeah, I love that part of my job. I love the part where we go out in the world and shoot a lot of video and then I struggle to put that script together. And so for me, it's where has that struggle been? If the struggle for me is like around doing administrative work and I don't know, budgets or insurance, reading through insurance plans, I'm okay with having AI do that. Right, or looking at contracts. They're actually quite good at that.

1:06:51Yeah. That's one thing. So I did just spend a lot of time talking to a lawyer. Oh, did you? Okay.

1:06:56Kara Swisher:Are there any lawyers in the room? So you're used to being a guinea pig for tech. Like often I like find out, you put something on, I'm like, oh, what is Joanna doing now? Like that, so you're used to that. This is your life. But talk about your wife and kids. How do their lives look different after a year of AI? And what experience tell you about the challenges in that regard? Are they thrilled you're done? My wife is sitting there. So I think she's thrilled we're done, mostly done with this book. The kids part, like I said, I did not know that this experiment was going to involve them so much.

1:07:29When I was pitching the book, I did not have them as central characters. But then I brought all of this home and it was very clear to me that this is much more of a story about the next generation because of the technologies affecting the next generation. We've seen that through both of our careers. And so I think they were affected positively, though now on Sunday nights when we have dinner, they both act like robots and they say we are cleaning robots and they initialize cleaning robot mode. Yeah, that's an old trick. I appreciate it. And so if my kids think that they need to be robots to clean the house, I'm fine with that.

1:08:04If it's caused them lasting damage, then fine. Yeah. But to the point of, I think they question AI now more. I mean, I have a four-year-old who really is talking about AI. He's constantly saying, Mom, are you going to work at AI today? Wow. I know. See, my four-year-old talks about his butt. That's really pretty. Mine too. They can get together. But, you know, it's his AI butt. It's nicely generated. But it certainly affected them, and I think in a positive way, but also they were exposed to a few technologies that I think are not right now ready in their lifetime, but will absolutely be ready in their lifetime.

1:08:48Right.

1:08:48Kara Swisher:So my very last question is from the audience here. What's the one thing we should know about AI? Well, it's an excellent question, audience members. Go ahead. It makes mistakes. AI makes mistakes. And we are going to let AI into more parts of our life. And we must be willing to ask and question the results. That's a really good one. Mine would be AI wishes it was as smart and interesting as humans. You think it wishes? You think it has feelings to wish that? I think it never will be. I think tech people, they always put the human brain as a computer. A computer wishes it was as smart and amazing as the human brain.

1:09:34Kara Swisher:Amazing. And I think we need to not let that get lost in them telling us how easy life is going to be. It'll be easy for them because they have money, and that's always been the same. And so use it when it's useful. Otherwise, I think we should rely a lot more on humanity than anything else. And I'm not doing that as a Luddite. It's just better answers. in the end. Anyway. Okay. This is a great book. She did a great job. Thank you so much. She did a great job with this interview, even though I'll pick a different moderator next time.

1:10:11Kara Swisher:Today's show was produced by Christian Castor-Russell, Michelle Aloy, Catherine Millsop, Megan Burney, and Kaylin Lynch. Nishat Kerwa is Vox Media's executive producer of podcasts. Special thanks to Rosemary Ho and Julia Sharp-Levine. Our engineers are Fernando Arruda and Rick Kwan, and our theme music is by Trachademics. If you're already following the show, as I said, you are not a robot. If not, you're yelling at Barbara. Go wherever you listen to podcasts, search for On with Kara Swisher, and hit follow. Thanks for listening to On with Kara Swisher from Podium Media, New York Magazine, the Vox Media Podcast Network, and us.

1:10:45Kara Swisher:We'll be back on Thursday with more.

1:11:17Kara Swisher:We'll be right back. like Solar Roller Shades from Blinds.com and save up to 45 % off during the Memorial Day Early Access Sale. Whether you want to DIY it or have a pro handle everything, we've got you. Free samples, real design experts, and zero pressure. Just help when you need it. Shop up to 45 % off site-wide right now during the Early Access Memorial Day Sale at Blinds.com. Rules and restrictions apply. Ryan Reynolds here from Mint Mobile with a message for everyone paying big wireless way too much. please for the love of everything good in this world stop with mint you can get premium wireless for just 15 a month of course if you enjoy overpaying no judgments but that's weird okay one judgment anyway give it a try at mintmobile.com slash switch up front payment of 45 for three-month plan equivalent to 15 per month required intro rate first three months only then full price plan options available taxes and fees extra see full terms at mintmobile.com

From the publisher

For her new book, “I Am Not a Robot,” journalist and longtime technology columnist Joanna Stern turned her life over to AI for one whole year. She let it fold her laundry, conduct interviews, drive her family around on vacation, review her mammograms — she even tried out a romantic relationship with a bot. Her experiences tell us a lot about the benefits — and limits — of this transformative technology.

Kara and Joanna talk about how Joanna went about weaving AI into her life, where it helped and where it didn’t, and the gap between Silicon Valley’s promises and the reality of using AI for everyday tasks. They also talk about the growing public backlash to AI and where people should draw personal lines around its use. 

Special thanks to The 92nd Street Y for hosting this live event. 

Questions? Comments? Email us at on@voxmedia.com or find us on YouTube, Instagram, TikTok, Threads, and Bluesky @onwithkaraswisher.
Learn more about your ad choices. Visit podcastchoices.com/adchoices

More from On with Kara Swisher

All 139 episodes
Joanna Stern Turned Her Life Over to AI For A Year — Here’s What She LearnedOn with Kara Swisher · 1 h 9 min
Listen in VO