Chat GPT’s new erotica mode

I'm Daniel James and you're listening to seven am. The creators of chat GPT have just made a surprising announcement. Erotica is coming to the world's most popular AI platform. It's the latest step in a transformation of the Internet, where artificial intelligence is shaping not just what we read or search, but how we think, feel and even form relationships.

So Sam Moltman, one of the co founders and the lead of open ai, announced that the organization is going to make some changes to its while we popular chat jipt product, which it released in twenty twenty two. He said, with reference to some mental health issues that have been linked to use of its chatbot, that the organization open ai had been successfully able to mitigate those.

There are fifteen thousand people a week that commit suicide, about ten percent of the world talking and charge of BET. That's like fifteen hundred people a week that are assuming this is right, that are talking to Chaga bet and still committing suicide. At the end of it, they probably talked about it and we probably didn't save their lives. Maybe we could have said something better, maybe we could have been more proactive.

And also because of a new approach to it where they are going to start age gating the chatbot and trying to figure out us ages and restricting certain functions so users, Sam Oltman said, it's time to let adults be adults, and so we're going to allow chat chipet the products that's used for everything from office productivity to search, to also write erotica for our adult users.

So we'll get to this idea of educating a little bit later on, but first, can you just lay out for me what open ais founder Sam Milton means when he talks about their previous attempts to deal with mental health issues.

Yeah.

One of the biggest stories over the last year or so with these AI chapots has been how AI chapots appeared to have been playing a role in fueling egging on people's mental health breakdowns. There's been examples of people killing themselves citing chat jipt as part of the involvement and involved in that has been this kind of often sickophantic or encouraging relationship from chapbot.

Like many teens, Adam Rain started using chat gpt for homework, which his parents Matt and Maria supported as their son as innocent questions like can you become a doctor with a biocam degree.

There's quite a high profile case on in the US at the moment where the parents of a child who who killed themselves are suing open Ai because they said your product actually helped him along this path by giving information about how to kill himself.

This morning, chat GBT hit with a scathing new lawsuit alleging it helped a sixteen year old boy die by suicide.

And so as a result, OpenAI has started to release features to kind of deal with these mental health problems, and so, you know, in August this year they released features that for example, they said that they were going to use their systems to try and detect where the people were spending long sessions on there not getting sleep, having manic episodes and rather than having the chatbots as they often want to do to kind of encourage people's

So chat jipt is part of all of this.

Is said, we're also going to make it possible so that, you know, you can have some oversight over how children are using this app, which is still available to anyone over the age of thirteen, And so Sam Moltman is saying, we've dealt with these mental health issues, We've got them under control, and so as a result, we can loosen up.

Our chat jipt to let adults be adults.

So they're changing the way the bot works to loosen the way users can engage with it. What's behind that loosening in your view cam?

Yeah, So, I mean, you know, I mentioned before this kind of dependency and some really severe mental health issues that had kind of come into the four with some high profile incidents, but there was this kind of broader thing where people just started observing that like that, a lot of people seem to be developing quite strong like dependence on earlier.

Versions of chat ChiPT.

A lot of that was linked around a previous version of chat Chipity, which was Chatchiput four O, which was very popular because it was notoriously sycophantic.

It was always saying, you know, whatever you would say.

Would be like, that's a great idea, you're very wise. May I suggest that you do X Y Z, which you know, if you think about it, For the average user, that might make you feel good. Maybe that keeps you a little bit on the app a little bit longer. That maybe means that you want to subscribe so you're

They were kind of.

Developing what people named ai psychosis, where they were talking about theories of recursion and quantum physics and saying that they'd uncovered secrets of the universe or new scientific discoveries when they actually hadn't.

And so this chatbot, which was notoriously.

Very friendly and very very affirming, had led them down this path of abnormal behavior.

So that was a previous chat chiput.

And so when open i released a new version in the middle of this year, it was supposed to be an improved version, but one big change that came with it was that the feel of this version was different. And so all of these people who have these very strong relationships with the previous chat rept seemed quite happy about it, because it's almost like their friend had just disappeared. There was this big kind of backlash, and ultimately Sam Oltman and the open ai team said, hey, we're actually

It to them.

Coming up, open ai claims it was founded to benefit humanity. Now it's transitioned to for profit. Cam let's talk about open ai. How has it changed as a company since I first started.

So this is a really really interesting part of this whole story.

Open ai was founded in twenty fifteen, and back then it really encapitulated this idea of AI safety.

So we have always viewed our kind of role on the world in our mission as we want to just increase the innovation output of the world, and that is good because it's intellectually interesting.

People who'd been interested in artificial intelligence had long spoken about, well, what happens when artificial intelligence gets really smart? What happens when it becomes capable enough to take over the world, What happens when it becomes capable enough to convince people that it is alive and they will form relationships with them. So they had long been an interest in saying, how can we not only develop this technology, but making sure that when we usher in whatever this new era is,

I think the only way that you kind of stay in the stable configuration is if you can have a lot more innovation and everyone's life can get better every year, and you have to growth that way.

So that's what we'd like to do.

And so that was seven years a century of the company not producing any consumer products. It was putting out research developing technologies until twenty twenty two when it released chat jipt at the end of the year.

You know, now the.

Most fastest growing consumer product of all time. It now has like eight hundred million users. It's got contracts with businesses, it has actually signed its first contract with the Australian federal government.

The government's job which we embrace, the responsibility that we embraces to find that middle path so that we can get all of the game changing benefits from artificial intelligence in our economy and in our society without dismissing or ignoring the risks including.

But at the same time, it's a company almost caught between its past and its future, where it was like we're going to do things safely, we're going to do things slowly, we're going to help the world. Now it's a very much a closed company that is producing a product with the intention of making money, and it is doing it in a way that it's releasing a lot of features that increasingly people who are following it are

And it's now a far more creative space with a lot more competitors tell me about how they might be impacting some of open AI's decisions.

So I mean, there's all kinds of competitors.

El Musk has his own competitor, Xai, which now owns the company forming known as Twitter, has this consumer AI product called Groc. Has very few safety guardrails. You know, it for a long time, has kind of done explicit content. In fact, a recent Groc version that was released a couple of months ago would produce explicit images of people.

These kind of relationships with our users.

And so these changes to chat GPT coming at a time when the whole Internet is changing. We're seeing social media age restrictions coming into place here in Australia and as you say, chat GPT will start age gating its users. What do you think the impact of all this will be.

I think it's worth pretty this in perspective that pretty much as long as we've had mass communications, some kinds of content has been restricted based on age. You know, we've had TV ratings and movie ratings, and you know they're often restricted.

You know, movie theaters would check your ID.

Or video shops would say, hey, I'm sorry, you shouldn't be you know, taking a copy home of basic instinct or whatever. The Internet was not really part of that for the last twenty five years for a variety of reasons, including the fact that for a long time it was kind of niche. But now, over the last few years we've kind of seen a growing push to say, hey, we actually just can't let kids do whatever on the Internet.

And cam as more and more people engage with chatbots on a regular basis, whether it be through more these more explicit developments or otherwise, how do you think it's going to reshape our human relationships.

What could possibly go wrong? What could possibly go wrong?

We're always alarmists, but I think we kind of have to be, because you know, I like to think about how technology can be used to help us. But we shouldn't be blindly optimistic because the companies that are releasing these products have certain incentives that aren't necessarily aligned with incentives of you and I.

You know, open Aye, which has.

A trillion dollars of bills lined up to run these services that it's promised, has to get people to use its products. It has to get people to subscribe to spend much more than they're already spending. You know, the incentives are obviously aligned for them to be like, well, how can we get people really valuing our products, perhaps

But here's the thing. I think what I always bring it back.

To is when you have eight hundred million people using a product, there are going to be people in that population who are vulnerable. And my question is, in the pursuit of maximizing profits to try and pay these enormous bills that they have lined up that they just need to continue existing, do those people become essentially a casualty?

To make their company exist. And I think that's a really, really dangerous place to be in.

Cameron. On that happy note, thanks so much for your time.

By Potua. Thank you.

Also in the news, Jamie Mellen has become the second female jockey to ever win the Melbourne Cup. Her win yesterday on half Yours at Flemington came ten years after Michelle Payne made history as the first woman to win. Mallam said she knew she was going to win and when she passed her husband Ben, who was also racing. Amit's rates have been left on holds at three point six percent. The Reserve Bank's decision not the lower rates comes after news inflation has jumped much higher than the