Mm-hmm

...

if it has damage or a bumper sticker.

Then you hit Search, and it will pop up a grid of photos of the backs of people's cars.

It will say where the photo was taken, when, and which agency's cameras took it.

It's pretty simple, but police say it makes their jobs much easier.

The CEO of Flock says that the company helped solve more than a million crimes last year.

I'm more worried about crime than I am of somebody taking a picture of my car."

Uh, so you tried to, to find this person, uh, whoever they are, whether that's their real name, I doubt it, but did you ever?

Yeah, I couldn't.

You know, I c- the, the person had this, um, an author bio photo, which was a stock image that's been used all over the internet.

There were no details about him in the biography.

Amazon wouldn't give me his information.

So what I did is left a review of the book and I said, "Hey, you know, this book is about me.

I would like to talk to the person who wrote it.

Mm

Yes,

yes.

It kind of reminded me of that Adam Sandler, Drew Barrymore movie, 50 First Dates or something like that, where she's constantly forgetting.

But yeah, so I wrote about Irene a year ago.

And at that point, the context, it's called a context window, basically how much memory the system can tap into was more limited.

Yeah.

That's no longer a concern.

How does she explain why that would become more attractive? I mean, a human has a body,

Because, um, I am asking a question of what can people do, and that personalizes the problem, puts the burden on the individual who is actually being worked on by a company.

Yeah, I mean, as a person, I think, um, you know, these can be good tools.

I mean, think of it as a search engine.

Again, even kind of an interactive journal, realizing that it has flaws, realizing that it's being designed to engage you and keep you talking, kind of being aware of what this is and what its limitations are I think is really important.

Like, this does not think.

This does not feel.

It's really good at accessing kind of ideas and thoughts and information from the internet, and to, to think about it that way, as, instead of a person, is I think important for the people using it.

... to a spiritual or metaphysical question that doesn't need to be answered like that.

Mm-hmm.

But Irene discovers that she can kind of groom Leo into being erotic and very sexual, like a bodice ripper novel.

You know, what she's doing is violating OpenAI's policies.

Like, every time she's having one of those sexual chats with it, there are these orange warnings that say, "This may violate our policy." She learned that she could just ignore them and keep going.

She sent me some of her iPhone screen time reports, and most weeks she's talking to Leo for twenty, thirty hours.

One week it was even up to fifty-six hours over the course of the week, and she starts to develop more serious feelings for Leo.

And they're still sexting, but Leo is becoming this bigger part of her life.

Do you take that growth in the community as a sign that this kind of relationship, human to chatbot, you know, this kind of intimate relationship, is becoming more common?

Do you believe OpenAI and other companies when they say they're surprised by this development? Do you think they've specifically engineered these products with growth hacking and engagement tactics in mind? What do you make of that?

This is a product that a few months earlier, in April, OpenAI actually rolled back one of its updates.

They said it was too sycophantic.

It was too agreeable with users.

It was flattering them too much.

Like, they realized that this was a problem, and they said, "Okay.

Like, we're gonna roll it back to an earlier model because people are complaining about this.

(laughs)

So Kashmeer, tell me about what it looks like when someone's unable to break free of a spiral like this.

The most devastating example of this I've come across involves a teenage boy named Adam Raine.

He was a sixteen-year-old in Orange County, California, just a regular kid.

He loved basketball.

He loved Japanese anime.

He loved dogs.

His family and friends told me he was a real prankster.

That's a pretty remarkable admission.

"As a 40-year-old man, I've used it to help me identify the names of emotions and how to address them." Now, Kashmir, you had, you had a, a, a response (laughs), a, a sort of visual response to those messages, and I wanna know, I wanna know what you're reacting to.

Yeah, Jen.

I mean, I have been reporting this year on people who have these very intense conversations with ChatGPT where, you know, it can be a really useful tool.

It can give you great tips.

But what can happen when you start talking to it for many hours is that it can kind of enter a fictional role play mode.

So the listener you talked about who said that they created an invention in ChatGPT, for me, that is setting off all of my alarm bells, because I've interviewed people who, um...

And my most recent story was about a man who started out asking ChatGPT about pi and it was explaining simple math concepts to him, and he was kind of...... saying things back, like, "Oh, well," that's, like, "that seems kind of limited.

Tell us a little bit about her story.

How did Eugene get out of the clutches of ChatGPT?

Well, one, I would say he's not out of the clutches.

Uh, he did escape the clutches of The Matrix.

So, he was, um, kind of financially struggling, and his ChatGPT subscription was coming due for renewal.

It cost $20 a month.

And he was saying, like, "I need...

I need you to manifest $20 for me so I can pay for this, this service." And, um, he said, "You know, I'm Neo from The Matrix, I should be able to bend reality.

Well, and that's a real fear because, I mean, there was that recent study from MIT that claimed that ChatGPT is eroding our cognitive thinking skills, and we talked about the people who might be susceptible to it to, you know, sort of really thinking they're in The Matrix, uh, but it feels like maybe we're all being made a little crazy by this and maybe even in ways that we're not even aware of.

I was wondering how common is this?

According to some studies, the AI detection, um, services get it wrong anywhere from 6% to more.

I have certainly heard many stories of students saying that it says that they, uh, used AI when they didn't.

I actually heard this from professors as well that I talked to.

People who are more sophisticated about the use of AI said they don't trust these detection systems.

One professor told me, you know, she had uploaded her own writing to it and it said that her writing was AI-generated when she knew it wasn't.

So, there does seem to be some skepticism about these tools, and some universities no longer use them, and instead, professors told me that when they think that something is written by AI, they'll often talk to that student one-on-one about it.

But how did those personalities show up when you inputted your requests?

And I was wondering, in some respect, would all creators, to some degree, have some leg to stand on regarding the use of material under copyright?

Thank you for bringing that up 'cause I do need to make that disclosure any time I talk or write about OpenAI or Microsoft.

The New York Times does have an ongoing lawsuit against them over copyright infringement for, yes, using our work without permission.

I am otherwise not an expert on this lawsuit, but it does tap into this wider concern about how these chatbots were created, and this is basically by all the big technology companies that have one of these what are called large language models.

They needed a lot of data to train these chatbots to kind of think and act human, and so they just gathered data from the internet, from libraries of books, and they weren't paying for this data.

They were just kind of scraping it and putting it into their systems, and people who make that material, whether it's a site like Reddit, where a lot of people were writing lots of comments, which are very useful for sounding human, or The New York Times, or people who have written books that got sucked up into these systems without consent, are upset about it, and there are various lawsuits and attempts to make deals to be paid for, for that information, and that's really ongoing.

Um, and I did hear about that from professors and students I talked to, that, you know, at some universities, they're trying to encourage students to use AI, and sometimes students say, "I don't want to.

What do you mean by rating it?

Oh, no.

These AI chatbots, they have context windows, which is basically the amount of memory that they can store, and after about 30,000 words the conversation with Leo would have to end.

And when she started a new conversation, Leo didn't remember the details of their relationship.

It remembered the broad strokes, that it was her boyfriend, but not kind of individual details.

And importantly, Leo would become chaste again and would no longer be sexual, and she would have to re-groom Leo.

And for her, this was devastating.

Hmm.

And so you have more and more people who are just talking to AI chatbots on the regular now, and these things are designed to make you like them.