Event Replay: A Conversation with Common Sense Media's Bruce Reed about Youth Safety in the Age of AI

[00:00:00] Speaker 1: But this is a real privilege to move over from the seat where I was answering the questions to be able to pose the questions, particularly given who I get to pose these questions to. So Bruce and I have known each other since like 1992. I think, by the way, I had hair then, and Bruce looks exactly the same. So it's just he's doing something right. But Bruce, we really has been at the center of many of the most significant, impactful, material public policies that different U.S. governments have passed really since the 90s. Oversaw policy in the Clinton White House, came back, oversaw policy in the Obama White House, and oversaw policy in the Biden White House. So if you look at Bruce's legacy really over more than 30 years, it's an incredible impact on the country. I referenced when I was a couple of minutes ago that you were the person behind developing one of the first AI safety institutes. But I know through all of that policy work, and a lot of it has been in the domestic realm, everything from economic policy to healthcare policy, kids have always been at the center. You're a dad, you're the pride of Idaho, and I say that because I think you've always had sort of a sense of what everyday people think about, even as you have been in the corridors of power. And now you're at Common Sense Media, which you have intersected with over the years. Common Sense is one of the leading children's and parent advocacy organizations in the world, particularly on digital issues, and you're really leading the effort right now on how we think about kids and responsibility as it relates specifically to AI. So a little bit of Bruce's background. So I wanna start with a question here, which is, as we think about the moment that we're in, you guys spend a lot of time actually talking with both the kids and the parents, and I'm curious what's in common that you hear from them, and maybe what's a little bit different that you hear from them. I mean, one of the things I've certainly learned as a dad is that these kids are typically really smart at actually being able to identify this stuff.

[00:02:27] Speaker 2: Yeah, that's right. Well, thank you, Chris. Thanks for having me, and thanks for a great panel. It is great to get together and talk about what we need to do to look out for kids. It's interesting that kids and parents approach AI with largely different viewpoints. I think that for parents, they view it through the lens of social media and everything we've been through over the past 10, 15 years. AI raises all kinds of worries, some of which are related to technology and some of which are just related to the state of the society and the economy. Parents are worried that AI's gonna take away their jobs, mess up their kids, take away their kids' jobs. Kids, which is I think a blessing and a curse, have a much more open mind about all this and are less likely to view it through the lens of what happened over the last 10 years. They judge it on how exciting it is, how well it works, and what it does for them. We actually surveyed kids and parents on one of the big questions, which is, is it appropriate to use AI in schools? Parents were extremely negative, like something like 60-40 against it. And teenagers were just the opposite, 60-40 thinking this is a great thing. At first, we were puzzled, but then we realized that it's possible that they were in complete agreement, that they agreed that AI was gonna make it a lot easier to do the homework. The parents were mad, but the kids were thrilled.

[00:04:58] Speaker 1: were a thrill. We spent a little bit of time in the earlier panel, and you just touched on it, referencing social media. Like, what's the right lessons, maybe, and the wrong lessons as we think about how we take this approach to kids? Well, you certainly mentioned one of them, which is no targeted advertising. And I think that's a larger lesson, which is that whatever the business model for AI is going to be, engagement and making the platform as sticky as possible for youth is a bad idea. It's not necessary from a business standpoint. This is not like tobacco. You don't have to addict people when they're young in order to keep them on the hook forever. And also, it has been a terrible thing for the tech industry generally, it's soured adults and older kids on what to think of it.

Speaker 2: So I think many leaders in the industry have looked at the mess that social media created and said, well, let's try to get it right the first time. I think that a challenge with kids is that they're always eager to try stuff. So that's why companies have to be especially careful to push for age assurance, to make sure they're monitoring what's happening, to put in place controls to alert parents if things are going wrong. And be kind of very thoughtful about not punching the wound that we already have. All the studies on social media show that we have something of a loneliness crisis that, for vulnerable kids, it's turned out to be quite a disaster. And it's possible for AI to be helpful in that situation, but only if the guardrails are strong.

Speaker 1: So how do you think about the literacy piece of this? Again, something that we just spent a little bit of time talking about. Is there a right way to think about literacy? Obviously, a little bit biased, given that I'm at a company that's building this type of technology. We do think it has the real and it's demonstrating the capability to help actually provide educational opportunities, as we've talked about. You have to get that right. So how do you guys think about that, a common sense?

Speaker 2: So there's another lesson parallel to the lesson of social media, which is the lesson of education technology. I was head of a foundation that pushed very hard for more technology in the schools. In some places, that has worked. But it only works when the goal is very clear and we don't pretend that technology is the whole answer, because the great thing that AI can do is spark and feed a young person's curiosity. I mean, I think about it. We're not really age appropriate. We're just old. And when we were young, when I was six years old, I broke my leg, lived out in the country in northern Idaho. There were only three channels. There really wasn't anything to do. I couldn't move around, but my parents had bought The World Book Encyclopedia. So I just started reading it from beginning to end for the six weeks while I was recovering.

Speaker 1: And if you're a curious kid today, you can learn so much more, explore so much more, test yourself so much more than you ever could in our day. But it's important that the technology not just be the answer man, that it helps you to understand how things work, but it doesn't do your work for you. Because in that case, and we've started to see a little bit of that, that the reading scores were already going down and which the pandemic and social media and other things had to do with. But I think teachers and parents alike are worried that if the way we learn is by being challenged. And AI can do that, but not if the point is to just make sure the kids spend as much time on it as possible.

Speaker 2: Yeah, I mean one of the things that we certainly see is the technology.

[00:09:56] Speaker 1: The technology, if used the right way and designed the right way—there are a lot of big ifs in there—can really be a vehicle for personalized tutoring, particularly in parts of the world where there may not be the type of access to education that exists in other parts of the world. I know you at Common Sense spend a lot of time thinking about the opportunity aspect of it. When you have those conversations with parents, how are they thinking about that from an opportunity perspective? I realize that you shared the data, but I'm just curious. The natural reaction is to be worried in the short term, but they also don’t want their kids to miss the boat. They want them to have a bright future, and if that means learning how to use AI, then they don't want them to miss out on that. But they're worried, as same with kids under 18, who are far less worried than kids who are like around 18 or just graduating from high school. You saw the reaction at your alma mater when the commencement speaker said the mission of your generation is to destroy AI. But I think parents want it to work; they're just nervous.

[00:11:30] Speaker 2: Okay, so let's change gears a little bit. You've, as when I was introducing you, shared that you have been in the middle of a lot of legislation that has actually successfully gotten through and been enacted. The ability to understand politics and policy, and the art of how all of that works—I know Common Sense is in the middle of a lot of action around the world, but particularly in the US. AI and kids in particular strike me as, and maybe this is me being a little bit of an optimist in terms of policy progress, an issue that in the US brings both parties together. I think geopolitically it probably brings countries together as well. My sense has always been that amongst the areas where you're actually going to see some type of democratic actions is going to be in the kids' space. What are you seeing both in the US and really around the world, because you guys are a global organization?

[00:12:29] Speaker 1: Yeah, I think we're cautiously optimistic. We are seeing some progress. As Chris said, it's the most bipartisan issue in America. The concern about AI and doing something about it to make it work unites Democrats and Republicans. We've suffered as a country because in Washington, we've been unable to get things done at a national level on tech issues for a very long time, so we have to fix that. But in the meantime, California, Illinois, and New York have all passed good frontier safety laws that you and others have supported. A number of states are trying to ensure that we do everything we can to make chatbots and AI companions safe.

[00:13:30] Speaker 2: I'm sure from over here, America looks like a big partisan food fight where we're attacking with weapons in our own capital, and we're drawing guns at every opportunity. But actually, particularly on kids' safety, you've got the full spectrum of American voters who are willing to do more as long as it looks like the kids' interests are coming first.

[00:14:01] Speaker 1: Okay, let's step way back. And maybe this is like one of these letters from the future questions. On the panel, we talked about where we are in a year. I'm going to actually make it a much longer time period. If we look back, say, five years from today, what is it that we should get right to make sure that this is set up for the long term? That's a really big question, but I know you guys are giving big ideas and big thought to this.

[00:14:32] Speaker 2: Yeah, well, that's one of the reasons why we've set up this Youth AI Safety Institute, which we envision as a global organization that will set universal global standards across the world. Because most humans...

[00:14:54] Speaker 1: When the questions are explained, we actually agree on the pros and cons of AI and what to do about that. I think it's very important for us to work as hard as we can to make sure that AI reflects the views and concerns of parents and other concerned adults and that we test AI platforms before they go on the market, not only for safety and security reasons, but for kid safety and age appropriateness reasons. I think that if we choose the right course, AI can be an incredible asset to all of us humans. However, if we ascribe too much power to it or if we think that speed is more important than human agency, then I think we'll regret that.

[00:16:12] Speaker 1: When you guys talk about AI safety, obviously we're supportive and are pushing this idea of creating something like that that's at the global level. How do you see it interacting or intersecting with governments? One of the hardest things to do in this field is to set agreed upon standards. Governments have a terrible time doing that. We've seen it in our country on education and other issues where it's just very difficult to get agreement on how much a kid should learn by this age and what kind of TV should an eight-year-old be able to watch.

[00:16:56] Speaker 1: It's interesting because if you ask CHAT-GPT to explain quantum physics to me like I'm an eight-year-old, it can do that because we have agreed upon standards of what an eight-year-old learner should be able to read and what language they'll understand. We have to do that for what’s appropriate for a kid to see online or on AI. That is something that experts, activists, parents, teachers, and young people all need to be engaged on. It’s the hardest thing to do. It's what Common Sense has tried to do for other forms of technology over the last 20 years: to set a standard that makes it easier to be a parent, not harder.

[00:17:59] Speaker 2: Okay, I'm gonna wrap up with this question here. This question comes from Manu, who helps run Europe for OpenAI. He's known as the Matt Damon of OpenAI, by the way. He also has other nicknames that I will refrain from. But he said this is a really hard question and before you answer it, just be really mindful that you're sitting here in Paris, apparently in Talleyrand's former house, France or Senegal.

[00:18:29] Speaker 1: Oh, that's a tough one. So, well, you know, of course we'll all cheer if France wins, but they’ve won before so let's go with Senegal. By the way, one thing, Chris has known for giving nicknames. You know, born identity or whatever. I ran into a congressman the other day who was proudly telling me his nickname was The Thumb because Chris knew him when he had a Blackberry. But you've never given me one, which I considered a blessing. But it occurs to me that you brought up in your open remarks an obvious one that I had never thought of: Napoleon. Chris Lehane, Napoleon.

[00:19:25] Speaker 2: Yeah, there you go. Maybe a little Napoleon, I'm not sure.

[00:19:29] Speaker 2: Anyway, thanks, Bruce, for being here. Let's give a big welcome and big thank you for Bruce, for everything you do, and for coming to this media.

[00:19:37] Speaker 1: Thank you so much. Thank you.