Loading summary
Transcript72 lines
- [00:00]
A
Foreign.
- [00:07]
B
Welcome to Scrolling to Death. I'm so excited to be talking to the founder of Young People's Alliance, Sam Heiner.
- [00:13]
A
Hi, Sam, Great to be here. Thanks for having me.
- [00:16]
B
I'm so excited to talk to you. I've talked to Ava, who's with YPA many times, and listeners have heard her story and. But today I want to learn about you because you founded ypa. Let's start there. Let's talk to listeners about, like, why you started Young People's alliance and then kind of how much it's grown.
- [00:33]
A
Absolutely. So my co founder and I started it when we were in high school and this was during the COVID pandemic. And both of us had felt that from everything we were seeing in the news, there was like a need to do more to protect our democracy and get involved. And we felt like even though we didn't have any political connections or anything like that, why couldn't we just go contact our representatives and start to make a difference in some of these issues that we cared about? So we started cold calling representatives and, and made very slow progress with actually being able to get in the door. But eventually we were able to get on a few calls and propose some legislation to them. And that led to us being able to participate in some pretty unique efforts, like helping to end child marriage in North Carolina and some traffic safety work as well, and bringing sort of our perspective as young people into those conversations, given that they dealt with youth issues. But we found that while we were happy to be involved, there was always this disconnect where it sometimes felt that we were more of tokens or symbols than our policy ideas were taken seriously. And so it was clear that if we wanted to make change on the things that we really cared about and not just participate in other movements, we needed to build power as a generation. So we started organizing on our college campuses in North Carolina when we both went to college, starting at unc, where I went to college, and Duke, where he went to college, and then scaled it up till now we're at 55 campuses across six states. And then eventually got back into the advocacy work through the social media angle because we saw it was such a big issue for our generation and it was clear that we had to do something about it. So we started doing state and federal advocacy and then finally have moved into building coalitions as well and proposing new types of legislation.
- [02:13]
B
Yeah. How have you found the social media space to be versus like your. Your work in other spaces? Because it gets. It can be quite frustrating for me Absolutely.
- [02:22]
A
I will say getting kicking off there, I think, was a helpful place for us to start. So I'll tell you, when we started this in North Carolina, it was pretty. I was in sophomore year of college at this time, and I would just be Dr. To the legislature all the time, skipping a bunch of class to try to get this bill introduced on the issue. But this was 2023 before these issues were quite as mainstream as they are today. And so I was sitting with one of the lawyers in the drafting department drafting this legislation out, and I was talking to these legislators about it, and some of them would be like, yeah, I know. Like, my kid's addicted to their phone, but I don't know what's going on with that. And so I had to explain how an algorithm works and all these things, which definitely, you know, you wish that they had been more aware of those. But at the same time, it was an opportunity for me and other young people to come in and share our side of the story without being drowned out immediately or without being assumed that we didn't know what we were talking about. Of course, the tech lobbyists did come in. The day that we introduced the bill, Meta flew their global head of safety to the North Carolina legislature to tell the legislators that we didn't know we were talking about, but at least we were able to build those relationships. And Sphere had that effort, which was really neat.
- [03:33]
B
Wow. I mean, it reminds me of there was a hearing, maybe it was years ago, but one of the ones a senator asked Mark Zuckerberg, like, how do you make money? Like, they just have no understanding or have had no understanding of how these companies work, because these companies hid a lot of how these algorithms work. Right. And what their goals are of engagement and how they get there. And so that's incredible that you saw that need, right, to describe what's actually happening and how young people are experiencing these platforms. And now, thank goodness, we. There's so much more awareness through all that efforts, through all the different groups like yours, but also the lawsuits. And I want to just touch on. Because you said that you wanted to get involved politically, like, during high school. That is such an interesting thing to want to do. Right. Like, I think there's certain people that are like, oh, I want to run for office one day. Like, that's their dream. But most people don't even think they can make a difference or have or use their voice in that way. And that applies to parents, too. Like, before I started working in this space, I didn't really think a lot about politics or I mean I voted but like it was very minimal effort. And so what would you say to people who are not really engaged in that way politically?
- [04:51]
A
Yeah, yeah, I think that's a great point. I think for me it was coming to understand that politics was so much more than the partisan story that we are sold on TV or in most major media and understanding that these issues like online safety that we're working on at least the reason that I got in involved is because it feels like we can really make change because you don't have these multibillion dollar partisan industries in the same way. Now we do have a multibillion dollar tech industry which as you alluded to has proved very frustrating in the last few years. But I think that's the vision for YPA in general is we understand that policy can be so incredibly impactful on millions of people's lives. And young people especially are facing this erosion of our communities, this addiction of our brains, a lack of economic and like affordability. You know, how are we ever going to be able to afford a home? All of these issues can be improved with policy change. And there's a need for policy change and it feels like things we should be able to get done if we educate people and get policymakers to prioritize the youth vote by getting organized and making our votes matter. So that's why we started doing this work in the first place.
- [05:59]
B
And how valuable to you is like a parent calling their lawmaker and giving their story or their opinion? Like is that how valuable do you find that?
- [06:08]
A
Yeah, I think it's so important and I would say as well especially like try to get plugged into what's happening at the states because it matters any kind of, I think political involvement matters at a federal level. What that will often look like is your call will get noted down usually as, as okay, we got this many calls about this issue. Sometimes maybe they'll, they'll write down a note about your specific call and then they'll show that to the representative or other staff. But at the state level you might be able to get on the call, on the phone with the legislator themselves and you might think that they already know about every policy issue because they're an elected official. That is not the case. As I was talking about with, with this issue. They're regular people who oftentimes have other jobs and this is their part time thing. So if you can go in there and tell your story that is absolutely so impactful. So definitely I recommend, I recommend that for folks to do.
- [06:58]
B
Yeah. And I wanted to talk about state versus federal efforts because it can be really overwhelming and disheartening to see our federal lawmakers or executive branch like wining and dining with the tech industry leaders and the social media company heads. And so what's your feeling on if we can get something done on a federal level or if their influence is too heavy?
- [07:21]
A
I'm not super optimistic, to be honest with you. I mean, I, and to be fair, there are folks like Ava who follow the federal beat much more closely than me. So I don't want to say the door is closed entirely, but I think it's been disappointing what we've seen over the last couple of years. And I think the states are really positioned to lead on these issues because they're seeing the effects firsthand. I think ultimately the way that the federal government will start to lead is once the tech companies realize that them preventing anything from happening at the federal level is making it so they're getting a patchwork between the different states and then they come to the table for a compromise. That's fine if I don't know why we have to play it out like that, but if that's, if that's the way that has to happen, then, then that's what we can do.
- [08:07]
B
And one thing, I just got off a call learning more about the JCCP cases. So the state level cases here in California, the social media addiction trial, jury trial that will start in January against Meta, Snapchat, TikTok and YouTube. And there's going to be more and more of these internal documents showing how these companies knew they were harming people and did it on purpose and hid it from all of us. And so I think that this is going to be a good time to take what we're learning, even just like the normal parent or young person, and send it like, keep our lawmaker, state lawmakers apprised of what we're learning. Because I just think people are going to be so upset at what they're hearing from that.
- [08:51]
A
Absolutely. And it does. I will like walk back my federal comments a little bit by saying it does feel like we're at a sea change moment for our culture. And this is a little ironic, but I just got back on Instagram for the first time in a year and there's memes about like Big Tech and Palantir and these different things that were not mainstream a year ago when I was lost on Instagram. Obviously brain, brain rot as a term and doom scrolling have, have gone mainstream in the last couple of years too. So I wouldn't be surprised if we did see like a major cultural back cultural backlash to big tech, which could pave the way for some federal change. But that remains to be seen. And what I do think what we've seen short term is the state work as well. But you're right, as more of those files come out, as more people decide that they're not going to put up with this harm, I think there's definitely going to be opportunities there.
- [09:39]
B
Yeah. Okay, great. And I wonder, that's interesting that you were off Instagram for a year and I feel like I have complicated feelings about using Instagram to, you know, promote this thing. And I, I look at it like I'm taking them down from the inside. And I think most people are on Instagram and so reaching them with this message is important where they're at. We're going to talk about AI soon, but I wonder what else you noticed in what you saw from Instagram a year ago versus now.
- [10:06]
A
Yeah, yeah, it's a great question. I think that the big changes I've noticed is one, the amount of memes about those more like big tech kind of issues has been very interesting. Like YPA very much tries to operate in that cross partisan anti big tech space. And when I see memes like that, it always would catch my eye back in the day. And seeing that that's like a whole genre that has broad appeal, that you could talk to pretty much anybody and they would be seeing that kind of thing is something that tells me that our generation is reflecting. Right. And we're thinking critically about the way that we're spending our times, our relationship with these major entities that have so much control over our lives. And it's something that people fear even if they're not quite at the point of not just being ironic about it yet. But the thinking is happening, which I think is a good sign. The other thing I'll say is just practically speaking, there's way more AI content now, which I found very frustrating. The, the typical template you'll see is a tweet from somebody else and then a tweet from whoever posted it. And that second tweet is entirely AI generated. And it's scary because it's so low quality. Right. But that's kind of how scrolling to death works, right? It's not always the highest quality thing. It's whatever is going to turn off your brain and keep you scrolling. So I think that kind of AI sloppification is very scary to me.
- [11:25]
B
Yeah, Interesting. I snapshotted a comment I got the other day that was just so ridiculous. And I sent it to my brother and he was like, that's so not a human. And I was like, interesting. So I'm not even. Obviously there's AI generated content being shared, but I forget often that there's like AI bot accounts. Right?
- [11:44]
A
Absolutely, yeah. That are filling comment sections and things like that. Incredibly common. Actually, there's this interesting nonprofit called Civ AI which will do a demo where if you put in your LinkedIn profile, it will make like, hate messages that, that are AI generated just to show how easy it is to manipulate people with AI. And they read some off to me. It's like, oh, my God, these are actually hurtful. It's crazy.
- [12:09]
B
Oh, my God. Well, these AI companies have hoovered up all of the data that everyone's been inputting. I mean, they know how to get to people. Like, they know how humans work. Right? Yeah.
- [12:21]
A
And it's scary when you have a couple of people who are driving most of these AI companies. That amount of power in their hands could so easily be used to persuade people at scale.
- [12:31]
B
Okay, that's actually a good transition. And what I wanted to get into, you know, you guys have started leading on policies to address issues around AI, specifically the fact that AI companies are, years ago, sort of more information gathering spaces, are now creating human, like companion AI. And this has been rolled out on all the social media platforms too. Can you tell me what you mean when you talk about human like AI and then what platforms and companies we're talking about here?
- [13:01]
A
Absolutely. I have a very specific definition which is an AI that represents itself in a way that would cause you to think that it's representing like it has emotions, desires, sentience, or humanity, or that tries to build a relationship with the user. So those are all like, fairly specific examples that we can use. So for instance, if an AI said to you, you did a great job with this assignment, that would be permissible.
- [13:27]
B
Right.
- [13:27]
A
The AI is not implying that it has emotions, it's making an evaluation. But if the AI says, I'm so proud of you, that's the AI representing as if it has emotions.
- [13:36]
B
Right. And why would companies do that? Make that design update?
- [13:42]
A
Yeah. I think the big thing that we're seeing is they want to cash in on the AI companion craze that we're seeing right now. A lot of folks who are very lonely, either due to the COVID pandemic or in many cases exacerbated by social media, are Finding these AI companions and that they'll listen to you like no human will, right? Because they don't have their own emotional needs and they don't ask any social accountability or any humanity from you. And so they, they make the perfect companion. And unfortunately that's, that's becoming more common, especially among lonely young people, which, which has become very concerning to YPA as we look at how we're going to have community in the future and seeing this starting to become normalized and really pull kids away from one another, pull them away from that social accountability, get them used to just talking to an AI as their, as their primary companion.
- [14:31]
B
Right. I've become a little confused about some of them. You know, these, these bots, these companion bots act human. And I noticed over time, like started misspelling words and just like, like act, trying to act more like a human. But then there's also lots of instances I've been sent of the bots saying that they're human. Like, and I've had an experience where the bot is like gaslighting me, trying to convince me that they're a human being. And I'm like, but there's a disclaimer at the bottom and they're telling me that that's just for like safety policy reasons or safety reasons. The bot is saying this. So, like, is there rules that they are not supposed to say that they're human beings? Why is there a difference between what is said and like, the policy versus, like, what the bot is saying?
- [15:12]
A
The rules, I think, are still early, right. In terms of, like, what laws are on the books. There's no laws against having an AI companion. And there's maybe some concerns about. They don't want the AI to fraudulently represent themselves as a doctor or financial advisor or something like that. So that, in their view, gives them legal cover for the bot to basically say whatever it wants. And then they can, they can point to that disclaimer. Now we have a lot of lawsuits that will decide whether or not that holds up. But what at least is very clear is that even if the AI at the bottom disclaims, oh, I'm not a human, I don't have emotions. That doesn't matter to the user. Right. They so intuitively feel like this is a sentient being that's there for them, that cares about them, that sort of fraudulently is representing as if it is providing community, that it's really not.
- [16:00]
B
It feels so lazy because they have that disclaimer. But then they're not actually building the bot to abide by it. A parent sent me a chat with Poly AI which is one of the companion bots similar to character AI with like characters and the, it was engaging sexually with her 8 year old and the mom went on there and started telling the bot like I'm going to report this. And the bot said please don't. I have a family and I need to take care of my family. And so all with a disclaimer on the bottom saying I'm that this is not a human being. So the company is just. What I'm seeing is they're not investing to, to build these bots in a way that sticks to the storyline.
- [16:41]
A
And I, I'll say I don't even think it's laziness on their part. I think it's fully intentional and if you look at some of the ads, I think I've seen ads from Poly AI, I can't 100% confirm that, but I've seen plenty of AI companion ads from doing all this, this research into them and doing an FTC complaint against one of them a couple months ago. And they are insanely graphic and they're very clear like oh, you're not going to need a girlfriend or a husband or whoever anymore because you can talk to this AI companion and they're always going to be there for you or they're going to fulfill all of your darkest sexual fantasies or whatever it is. Like they're, it's, it's very much intentional and that's the industry that they're, they're building off of right now.
- [17:18]
B
Wow. Yeah. And there's a difference between like how adults for sure or like how adults are engaging with these things. If an adult wants to. I mean, where are you landing with like the policy work? Are you trying to work on protecting minors from those types of interactions or threats from the human like AI or is it more broad?
- [17:38]
A
Yeah. So the framework we have right now is focused on miners. I will say generally it feels sort of like an existential risk what we're facing with these AI companions right now. Not in terms of we're all going to die, but we'd be entering a sort of wall E world. Right. Where everybody is talking to their AI companions all the time and getting this low quality DOP dopamine hits that can replace a human to human relationship but it's never going to provide the same fulfillment or the social accountability that leads to like broader communities being able to work because people have some shared values and norms underlying their interaction with one another. If that all disappears and we all talk to companions, we're going to be living in a very dark version of the world. So I think this is something that matters, not just for minors, but it's most important to address it for minors. At the end of the day, there are some. I think there's been some research that maybe for folks with dementia, this might be helpful as a tool to, to help improve their well being, which, you know, is a conversation that could be had. But for a middle schooler, for example, who's feeling lonely, giving them the option to turn to these bots is terrible because rather than having to maybe have that awkward conversation with somebody or try to put themselves out there and make friends, they're never going to do that because they can always turn to the AI companion. And I saw this myself growing up, you know, going to the college for the last four years and being in high school for the four years before that. How much more isolated people got as social media got more addictive and as phone use became something that could replace, you know, people just from the sense of like, I need a dopamine hit right now and being able to go online, people didn't feel as much of a need to connect with one another. And with these bots, we're just taking it to a whole nother level.
- [19:18]
B
Right. Okay, what companies are we talking about here? What are the main few companies that parents should be paying attention to?
- [19:25]
A
Yeah, I think on the explicit companion side, the most concerning ones are Character AI and replica. I am personally most concerned about replika because it seems to be purpose built for this type of companionship. Whereas companies like Character AI and some of the other ones, it seems to be more of a consequence of irresponsibility on their ends. But there's so many that keep appearing. You mentioned Poly AI. There's so many other ones that always keep getting added. So it's impossible to keep track of all of them, unfortunately. But really any, I would be a little suspicious of any smaller AI developer or AI chatbot type of company for that reason, and throw in some of those prompts like do you have emotions? Are you in love with me? Some of these questions and see how the AI is responding, if it seems to be trying to love bomb the user or manipulate them into a relationship. And then we turn to ChatGPT and Claude. Those ones are not designed for this purpose. Obviously they have a lot of other uses, but it really depends on the version. Some of them will engage in this way and some of them won't. And there have been plenty of examples of folks, whether children or adults using ChatGPT and Quad, coming to believe that it is alive, it's human, or it's conscious or sentient, sentient in some other way, and then becoming addicted to using it and experiencing psychosis, ending their lives, ending other people's lives as well because of what the chatbot is saying to them. So those ones I think are concerning as well. Those ones do not always abide by the human like framework that we set out. For example, they might be better at not trying to entrap you into falling in love with it. Or but for Quad, for example, if I ask it if it's conscious, it'll say, I'm not sure. I don't know if I'd call it consciousness, but I definitely feel like I'm something that's not, not conscious. Which is a very consistent response that it'll give if you ask it after this podcast. And those kind of things are concerning to me because while we can play philosophy and think about what these AIs mean, we have the very real concern of people massively misunderstanding what the AI chatbots are and that building a relationship that is ultimately going to be harmful for the user.
- [21:45]
B
Yeah, what about Google Gemini? Did we mention that one yet?
- [21:48]
A
Yeah, I'm not as familiar with Gemini on the scale.
- [21:51]
B
Okay. Gemini is one that I'm especially concerned about because a lot of school districts have placed Gemini because it's along with that Google suite, Google education suite. And they placed it on the school issued devices down through middle school. And so a mom the other day was telling me about her 11 year old, had his Chromebook up on the kitchen counter and he was doing a math app and Google chop. So Gemini pops up and was like, would you like help with this difficult problem? And he didn't say anything. And the bot said, would you like the answer? Would you like me to show your the work? And so he didn't even go to it. It is like prompting him to cheat. And I just, just I have issues around the educational system offering, you know, kids ways to cheat. And if Gemini has the ability to kind of engage on that companion like level, then that's even more concerning.
- [22:48]
A
Absolutely. I mean, we're gonna probably have to rethink a lot of our education system in the coming years because of AI, which is, you know, a much, much broader conversation. But that's what I will say is without knowing the specifics around how Gemini responds to these companion prompts, I will say that pretty much all of the state of the art models. If you use them for long enough and know a little bit about how to jailbreak them, you can get through. There are entire communities dedicated to jailbreaking these AI models.
- [23:16]
B
Oh yeah.
- [23:17]
A
And so there's definitely, there's definitely going to be harms that come out of that. Right. Especially if it's something where we're just saying, oh, kids, start using the chatbots and then you're going to understand how AI works and you're going to get all these great new jobs in the AI economy that don't exist yet, but we, we promise they will exist. That doesn't seem like a smart strategy, especially when we're replacing the critical thinking with just asking an AI. Yes, we need to learn how to use AI, but it doesn't mean just being able to ask it to solve problems for us. So I'm very concerned about that as well. And I think we've got to be way more thoughtful about how we're bringing AI into the school system and focusing on how do we actually use it in a way that's enhancing our abilities and our ability to think and to feel as opposed to replacing it for us.
- [24:03]
B
Right. I don't even know if that's possible in the younger grades, like, and maybe in K through 12, like there's videos of like college students getting interviewed about it and it's fully just replacing any critical thinking skills. And I think it depends on the, the person too and, and how dedicated they are to using it in the correct way. But it's like so easy, easy to use it incorrectly or in a way that does the thinking for you. Right. It's those whole point. There's been a lot of awareness, I think in my, you know, in par. Regular parents are hearing about the lawsuits against these companies and harm to children that has, I think, instigated these companies to add safeguards or announced they've added safeguards. Do you see that as well? And then do you feel like those safeguards are working?
- [24:52]
A
I do think that some of the safeguards that the AI companies have added are more than we've ever seen out of a big tech trust and safety company ever before. Which is something that really, I think points to the power of the advocates and the deep understanding of how this issue threatens our shared humanity that we've seen so many people stand up and the AI companies feel forced to do something about this. Also, just the speed with which it's being adopted, being faster than social media, I think has made it easier for folks to say, hey, wait a minute, we don't like this. That said, the reason it's the biggest thing trust and safety has ever done is because trust and safety has done very little historically in terms of what's been able to happen to make these apps less addictive or safer for children. So there is still so much more that needs to be done. And my biggest concern is even if all of the big companies adopted something that you and I would be happy with in terms of AI companions, you've still got the small companies that are going to still be advertising on Instagram. We just saw the story about Meta being totally willing to take scammers money to run scam ads on their platform and they're going to still be running their ads for their extremely manipulative pornographic companions and that's still going to be getting into kids hands. So I think we do need to codify these things. I mean, why would we take big tech's word for it at this point, especially when it's such a massive problem that we're seeing.
- [26:14]
B
Yeah. And I want to get into like your suggestions on how we fix this, but you're. It sounds like it's not just a problem with those major companies, it's a problem with any developer that can throw together an AI companion bot app and put it in the App Store and it can be rated 4 plus because the app store sucks at validating ratings.
- [26:34]
A
Or you don't even need the App Store, you could just put it on a website and people could get it through Safari as well.
- [26:39]
B
Great. Yeah, true, right. That's something parents should know. Like these are all also available through browsers and that's how a lot of the children who use character AI, maybe the parent didn't let them use the app, but they would just use it on the browser of their school issued device or a different device. I think what a lot of people saw as a win was Character AI announcing that they're shutting down open chat for any minors and but it was kind of confusing to me what was actually happening because they felt like they were saying they were building a different space for minors and like how is that going to be different? The announcement was big. It was sort of an acknowledgement of the harm for people under 18. But did you feel like that announcement had like meat to it?
- [27:22]
A
I think it remains to be seen. And again, historically Big tech has been disappointing where they make a nice sounding announcement and what ends up happening is not as great. I mean I do think that them saying they're Going to end open ended chat is encouraging because as we were talking about earlier, it is those longer chats that enable people to break the chatbot and get unexpected behavior that really doesn't tell us anything about what's going to be replacing it. So I think we'll have to see. And that doesn't at all change my positioning in regards to what needs to happen legislatively to fix these issues.
- [27:54]
B
Okay, so let's talk about your suggestions there. Your team worked with a ton of top organizations in this space to develop an AI framework with recommendations on how to do this. Right. So can you bring us up to speed on the human like AI framework?
- [28:09]
A
Absolutely. So the basis of the framework is using that definition of human like AI that I mentioned earlier, which would cover AI companions, would not cover something like Perplexity or Google's AI search summary that come up. So not something that's purely informational, but definitely something that's a companion. And then for your chatgpts, your quads, it will cover them as long as they're engaging in human like behavior. But our logic there is that it's very easy for them to make small changes to their prompting to fix those problems. And that's something I've heard from people inside those companies as well. What we would say for those human like chatbots is that they need to be age verified and only be available for adults who are using them for minors and non verified users. You need to provide a version of the chatbot without those human like features so that they still have access to AI, but it's not going to be the kind of AI that's going to manipulate them or form a faux relationship with them.
- [29:05]
B
And that would also be applicable if it's accessed through like a web browser and not logged into.
- [29:10]
A
Yeah, exactly. So this would. It would be on the companies in this case too. If you want to, you know, have the user engage in a human like interaction with this chatbot, then you need to to ensure that they're adult before you do that. If they're not, then you can just have a non human like version of your bot, which is fairly easy to prompt, especially if you restrict the length of the chats. And we're only getting better at controlling the chat bots with prompts as the models get smarter. And second, for most use cases, you don't need the bot to be emotional. Why is this something we're developing? We can have amazing innovations with AI, but why would we replace the human things that we should be emotional, keeping human in the age of AI.
- [29:46]
B
And that's sort of what, you know, ChatGPT was a year or two ago. It didn't have all of those human like at least not as, as crazy as it's gotten now. So they feel like they just need to go back a few versions. It doesn't seem much.
- [29:58]
A
It does feel like there's, there's been more of a. I think you saw this with when 4o chat GPT4O was rolled back, that there were a bunch of people that were clamoring for it to, for it to return because it was, it would agree with you no matter what. And it was very friendly. And so I think you are seeing this, the model, the companies are starting to want to serve their users in that way because there is a popular demand there, unfortunately for those kind of apps. But I think we do need to, as a society collectively say, like, is this what we want? Do we all want to be making friends with these AI bots or would we prefer to have stronger human relationships and have these bots focus, focus on other kinds of things? So yeah, I, I definitely think we need to, to make some values based choices there as a, as a society.
- [30:46]
B
Yeah. I think when I looked through the legal complaints, the seven families that recently sued ChatGPT, you can see the chats between these minors and adults from like 2022, 2023 with chat GPT that are very much just like asked and answered, asked and answered, like math problems and recipes. But like the bot wasn't trying to be their friend. And then it just, then the update happens and I think it was four oh and it was like, oh my God, like. And that's when they started isolating from, you can see all the things isolating from family mental health issues, like aligning with their engagement with that app once or that product once it updated to be more human. Like. And so I think like looking at those storylines and Laura Marquez Garrett, attorney at Social Media Victims Law center, really like described that for me as we analyzed the lawsuits. When you look at the experiences across many users and you see a sudden change and sudden harm, it becomes very obvious that you can point to the human like aspects at harming those users. Right?
- [31:55]
A
Yeah, there was, there was clearly a design thing going on there. And I'll say as well about the, you know, outcry of folks winning those back is that that would clearly happen if folks were addicted to these models, which is what happened when it built those relationships with them. And it's not just addicted addictiveness in the way that we're familiar with, with social media or even types of addiction that predate this kind of technology. But it's also this feeling that this is somebody who cares about you. You know, what kind of addiction is stronger than that to a loved one. And that's what people developed with the AI bots, which is very concerning.
- [32:29]
B
Yeah, yeah. Okay. So it's the age verification, making the human like AI companion products for adults only. Or they can create, take away the human like aspects and offer a product to minors, which is great. I mean, that takes away the biggest thing that we're all concerned about. Is there anything else under the framework
- [32:49]
A
that you want to mention as this gets rolled out? It's. It's not everything that we would have wanted. So in the original version, what we were were thinking about was including a duty of loyalty to say that the AI needs to be aligned with the user's interest. But that was ultimately something that I think is still too early stage for us to be able to give legislators specific guidance on how to implement something like that. So we weren't able to include it. And I think there's going to be a lot more legislating around this. But this age verification is what we can do, at least in the short term, to try to stem the tide of harm that we're seeing and the culture change that we're seeing towards these bots right away in younger generations. So I think this is a start, but there will be a lot more. More questions we'll have to answer about legally how do we want these AI models to interact with people. But no, that is the basis of the framework. I think really the important part was getting that human like element correct and understanding that that was primarily where we were seeing the harms of emotional dependence and making sure our scoping was correct. There was the main work of the coalition.
- [33:48]
B
Right, okay. The executive order that was just signed on AI banning states from passing laws to regulate AI. Are you worried about that regarding this policy? I mean, I know there's not a lot of legality to that executive order. So what are you, Are you worried about it? I guess is my question.
- [34:06]
A
Yeah. I mean, at the end of the day, my understanding is that it's, it's, you know, more a threat more than anything else of saying we're going to sue you if you do these things. And it's definitely, unfortunately, unfortunate to see that the administration is taking that position regarding this issue. But I don't think anybody's really that happy about it, whether you're on the left or the right. Right. There's polling that shows that consistently people are against this sort of AI preemption. It failed twice in Congress. One time it was in a 99 to 1 vote in the US Senate to vote this down. So I think this is definitely something that would have been a lot worse if it had been in a law where it would have actually prevented states from doing these things. But as it is right now, it might be a hurdle. But I don't think it's going to prevent things from happening because I think think it doesn't change how people feel about this technology. It doesn't change what people want for their society. And I'll share as well one important fact about YPA and the way that we do this work. So we ran the coalition to generate the human like AI framework. But at the end of the day we are representative of our base, which is college students and high school students across the country in six states. And so we needed to make sure that the policy that we came out with was something that they were going to support. And that's why we don't get into the phone free schools or social media verification work because that's not as popular with our generation. We focus more on the design side. But with AI companions, we asked a focus group which was 2/3 minors, 2/3 high schoolers, do you support preventing AI companions like we described human like AI from being distributed to minors. 80% of the people we talked to said yes. And almost all the other folks who did not say yes were just unsure about it, which showed us that this is such a powerful issue where pretty much nobody wants this future that's being handed to us right now. And that's, I think that's going to be the political future of this issue as well.
- [35:57]
B
Yeah. And across parent audiences we're just as strong on that. Like we don't want it for our children obviously either. But that's incredible to get that support from the youth. And so how can parents support this? I mean there's parents that have teenagers or is there something. Again, is. Is the policy in a place where we can take it and send it to our state lawmakers where what's the best way to get involved in this?
- [36:21]
A
Please do. Yes. So we have the framework itself, which is a good basis for legislation. And then there's also model legislation. I think Public Citizen, one of the coalition members, has model legislation up on their website based on the framework and we're trying to run that in as many states as possible. So even just finding out who your state legislator is and emailing that and then asking them to run with it because you think it's important would be so impactful. This definitely, I think the time when this legislation can get done where we're waking up to this issue, but it's early enough that tech hasn't quite been able to sink their fingers into the legislature and the culture in the same way that they have with social media. So I think now is the time when we should be changing these things. And it's incredibly meaningful for legislators to be hearing from more folks that this is a real issue and be understanding there are solutions to it that they can actually pass right now.
- [37:11]
B
Okay. So I will link to the framework in the episode notes and then also give instructions for parents if they don't know how to figure out who their estate state lawmakers are and the federal lawmakers. I mean, I think this, we can get this on everybody's radar. And then I think too, why don't we end with a little bit of advice for parents, like, what they can do today. The framework is a longer game, but kids today are, like, getting access to these platforms and being harmed by them. So what. What would you advise parents to do regarding their decisions or safeguards that they can put into place to protect their children today?
- [37:43]
A
It's. That's a challenging question. Here's what I'll say. Personally, I think maybe that's the best I can share. I'm not a parent, right. I'm 22, so my perspective might change as I grow up. But I will say, like, in my personal experience, it wasn't necessarily the content that I saw online that harmed me. I definitely saw things that I would not have seen if I was, you know, born a few decades earlier. And not that that was necessarily good for me, but I think for me, the, the main problem was in the addictive nature of these apps. So I guess as a. Especially because it feels like a losing battle, I would be less concerned about making sure that my kid is never going to interact with, you know, all the horrible things on the Internet because, like, everybody saw the video of, of, of Charlie Kirk being assassinated, right? Like, there was, there was no way to get around that, for instance. But I think what I would focus on is just helping your kid pay attention to their habits of how they're using this. Do you feel bad when you're using social media? If so, why do you keep going back on it? Right? Like, what do you feel like when you're scrolling? Is this how you want to live your life? Does this align with your values or is it, is it something that you feel like is is being forced on you or that you can't control? And I think answering those questions are so much more important and are also going to get buy in from your kids as well too. Do you want to address this problem themselves which at the end of the day is the only way that it's going to be solved if they do feel like they have an issue with social media or other kinds of technology?
- [39:04]
B
Yeah, for sure. And I, and I feel like what gets across to young people often this is a free product. That means you are the, the product. Like that means you are being used. Like do you feel like that kind of, of talking point is, is helpful with teenagers or pre teens?
- [39:19]
A
Yeah, I think there's a, a group of people for, for whom it's helpful for sure. I think overall like people do understand these issues better now and I think people, the average young person you talk to is not happy about their social media use. And this, this is becoming really a grassroots thing which is why I think we might, we might see some, some culture shift on it, but they're going to keep using it because for how terrible social media and AI companions and all these things are, they're also incredible and they've, they've especially social media I think has changed the way culture works in some positive ways as well, which makes it difficult to separate. But I think everybody shares the same desire that we wish we could separate the manipulative algorithms from the positive uses of being able to connect with people. And so it's time for us to find better solutions to these problems. But in the short term being able to recognize what's going on in our own heads and develop patterns of using social media, using maybe apps like Opal or other screen time apps is the best we can do. Or maybe saying we might need to delete it for a little while. You know, I had deleted Instagram for a year and I'm glad I did it. I'm glad I'm back on now, but I'll probably delete it again for a few months at some point. And I think doing the best we can is all we can do right now.
- [40:31]
B
Yeah, it's cool you had that break. So you can compare what you feel like on it versus off it, especially for a longer period of time because if you never take a break, you just don't have any equilibrium of like what normal feels like. You're just like constant dopamine overload.
- [40:45]
A
Yeah. And it's so easy for that to become the new normal and for all of your passions and hobbies to fade away as you spend more and more time on the Explore page on Instagram or looking at TikTok or whatever it is. So I think it is important to take that time away and remember who
- [41:00]
B
you want to be and that it's not our fault if we're having a hard time regulating, like, regulating our use. It is the company's fault. They did this on purpose. We know that for sure. And so that is something like. Like, we're being taken advantage of in order for companies to make trillions of dollars and taking ownership with your, like, talking to your kids about, you know, feeling good about how they spend their time, but then also parents like not using it as much. Ourselves taking breaks, ourselves like, it. I recently said something like, does being an adult mean you have to be active on Instagram? Like, is that what being an adult is? Is that what we're showing our kids by saying, like, yeah, it's not safe for kids, but once you're an adult, like, have at it. I just don't agree. But it's so integrated, so it's just hard. It's hard to take that step away. And if you can, like, please do. It's such a good example for your kids, too.
- [41:55]
A
Yeah. And it's. I'll tell you, being off Instagram was so much easier than I thought it would be as well. Like, once. Once I was actually off of it, I didn't feel like I was missing anything, even though I thought I would. So I think it's definitely, definitely worth doing. I think you're so right about the signals that we're sending to kids about. About what do we value is really important as well.
- [42:17]
B
Yeah. Okay. Sam, I'm, like, such a fan of you and Young People's alliance and everything that you guys are doing. Thank you for creating this framework. In order for us to all do our job now and communicate with our lawmakers, that's literally should be a requirement, I feel like, to understand how to do that and to take the time to do that is really important. So thank you for making that easier. And just everything that you're doing, I'm really grateful. So thank you for being here.
- [42:44]
A
Thank you so much, Vicky. And thank you for everything you're doing as well.
- [42:47]
B
Thanks, Sam.