Fertility rates in the U.S. have fallen by 22% since 2007. This also happens to be the year the iPhone was introduced to the world by Apple.
The less time people spend connecting in person, the less likely people are to have babies. But the question is can the reasoning really be that simple?
The motto for justice in democratic countries is “innocent until proven guilty,” but what if it were the other way around?
Algorithms and data analytics being used in policing raise troubling concerns about privacy, autonomy, and freedom.
Guests:
- Hernan J. Moscoso Boed, macro economist/associate professor of economics, University of Cincinnati
- Matt Burgess, senior writer, WIRED
- Russell Holly, director of commerce content, CNET
Transcript
This transcript is generated with AI. To ensure its accuracy, review the audio file.
Amy Juravich: Welcome to Tech Tuesday from All Sides with Amy Juravich, a show where we share stories about science, technology, and the future of our environment. Fertility rates in the US have fallen by 22% since 2007. This also happens to be the year the iPhone was introduced to the world by Apple. Research suggests that the iPhone might have had a pool in the drop of fertility rates. The less time people spend connecting in person, the less likely people are to have babies. But the question is, can the reasoning really be that simple? Joining us now is Hernan Moscoso-Boed, microeconomist and associate professor of economics at the University of Cincinnati. Welcome to All Sides, Hernan.
Hernan J. Moscoso Boed: Thank you for having me, Amy.
Juravich: So, how did you test that the iPhone specifically caused a drop in birth rates? Tell me about the study.
Moscoso Boed: So we split the population in different ways and the group that is easier to analyze are teenagers. This is a group of people that for the most part don't want to have kids. And we see a very sharp drop in their fertility after 2007. You mentioned that there was about 20% fertility drop for the whole population, but for the teenage, uh case it's about 70 percent and so that's uh much more if you will in your face the effect and so then um we looked at time spent socializing and so these teams um shifted towards screen time um they're what we call the unstructured time they still go to practice they still go to school, but when they are free to do whatever they want.
Without any adults supervising, they choose to spend much more time on the screen. And that's what we see. So there's much more times individually on a screen. And then behaviors that are connected to being in person end up being affected. We go for fertility because it's very well measured, but other behaviors also show this impact. And we do look at those in the team study. So, for example, so we see less time in person, fertility drops very fast, but also we see an increase in teenage suicide increasing with the isolation of this community. Other behaviors like violent crime at teenagers, it's also dropping.
So that's basically the main story. So, more people having the phones. The social interactions moving from the physical world to the digital world, isolation kicking in, and then behaviors that are connected to isolation, and one of them is fertility, ends up being affected. And then we look at that. Yeah, go ahead.
Juravich: Well, you kind of said at the beginning, though, that teenagers would be a sector of the population that maybe do not want to have babies, right? So in that aspect, the iPhone helped with keeping people apart and they're on the screens and so therefore less teen pregnancy, right. Is that what you're finding, that the iPhone helps in that area? Is that a positive, I guess? Yeah.
Moscoso Boed: Um, so as an economist, we want to stay away from goods and bads, which we're mostly describing, but you're right. Yeah. Um, uh, having fewer 17 year old girls pregnant for the most part, good, um, now the suicide story is not, we can agree that that's, and so what we're finding is that the phone is changing how we interact. So humans have been, we are a social animal. We've evolved over 200,000 years. And in the last 20, we've been interacting differently in ways that we can also measure. And so that's the gist of our story, more than whether it's good or bad.
After this work was shown around the world, people have been saying, well, but it's great that we have fewer kids, people calling for the environment, maybe we're not gonna be putting so much pressure on the environment. Teenage pregnancies falling, I agree. There's also a flip side, fewer kids mean fewer workers in the future. There's an unbalance between retirees and workers, social security systems might be stressed. So when we go into whether this is good or bad, I agree that this is not all good or all bad. It's a challenge. But the way we look at it is that Well, this social animal that we call a human is interacting differently now that we have the screens on our hands.
Juravich: Now, in 2007, when the iPhone first came out, I can't think that there were many 17-year-olds that had them. I mean, we grew into everyone having a smartphone. It wasn't just right away in 2007. So did you see more of this in more recent data rather than in 2007?
Moscoso Boed: So since the beginning of the 21st century, in the case of teenagers, fertility was relatively flat. And then in 2007, we see a kink and it starts to go down, down, of course, the difference between 2006 and 2008 was not very dramatic, but when you go to today, that's a huge difference. So yes, more and more of our lives went digitally and then we started to see that impact grew over time. Of course, yeah.
Juravich: Did you see this trend all over the United States? Were you studying just particular cities, or I don't even know, were you studying globally? Yeah.
Moscoso Boed: So we did both. So we looked globally and globally is very interesting because we see when the digital ecosystem arrives in a country, right? So think about a developing country that receives the digital revolution later than the U.S. And what we usually see is that their drop in fertility happens later. That's what we see around the world. And it happens in very different countries. So I'm from Argentina. It happened like about five years after the US. But it also happens in the Middle East, it happened in Western Europe.
For many different reasons, people talk about other reasons in each country, but you give them the phone and the acceleration happens, the decline in the fatigue. And then we went across the US, we have county level data about the rollout of 4G. And so we know what percentage of each county had 4G coverage in each. Yeah and we can interact that with two things. First, the fertility story, the numbers on fertility, and also what we use as what we call instrumental variable, which is the randomness generator, which is ruggedness.
So we have counties that are more rugged than others that end up being more expensive to install the infrastructure and they get the 4G network later. And we observe that because of that, fertility is affected late. That's when we start talking about causation instead of correlation.
Juravich: This is All Sides on 89.7 NPR News. We're talking about iPhones and the drop in the fertility rate with Hernan Moscoso Boed, Associate Professor of Economics at the University of Cincinnati. Now, the attitude toward teen pregnancy or having a child in early adulthood, maybe under the age of 20 or something like that, has changed drastically since maybe our grandparents' time. You know, decades and decades ago, it wasn't as. It wasn't seen as negative for someone to have a baby very young. So could societal expectation be another shift in the declining birth rates among younger people or is it all the technology?
Moscoso Boed: No, no, it could be, I think the story is going towards something like the modernity is being accelerated with the access to digital technology. All those forces are there, you are right, but then we saw the kink in 2007, right? Those stories about early motherhood were true in the late 1990s. But nothing happened there. And then we had to, once we were given the digital ecosystem, and we put the date in 2007, but there's nothing special about the iPhone. It's mostly about the whole ecosystem and the digital world, right? So the social media revolution, all that stuff that is attracting our attention into that.
Juravich: Okay, and but what about the the social media that existed before the smartphone before 2007, because we all spent hours sitting at our computers, whether they be desktops or laptops, chatting on instant messenger, looking at my space or Facebook on the desktop computer. Wouldn't that have had a similar rate effect on birth rates, especially for younger generations, just just AOL instant messenger even?
Moscoso Boed: Theoretically, yes, the data is showing that it was not as.
Juravich: Okay, so having the technology in your hand is different than a computer?
Moscoso Boed: Yes, absolutely, and it's, we have a much deeper access to, so what we access through the phone in 2026, it's much more complex and deep than what we did in 1998. So even if you had given me a phone with access to that rudimentary social media, probably it was still not deep enough to cause all this effects that we see. Um, the other thing that, uh, we did is also look at the, the people that want to have kids. That's also different.
And so that we have the teenagers that don't want to have kids and they end up not having the kids because they spend less time together. What about with the older adults? And probably this is the time where the babies are more wanted and so the people that want have a kid, they do go through a lot to get a kid to get pregnant. And what we find is now the digital world is distracting us. So we call it wide and shallow relationships. Now we have access to this ability to have wide and shadow relationship in social media, for example.
And where does that attention and effort coming from? Well, it's coming from not developing a deep relationship that end up producing family and fertility in kids. And that's a different story from the older Kohol.
Juravich: Yeah. Okay. So can you infer that someone who has higher screen time might be likely to have less children?
Moscoso Boed: That's basically what the, yes. So what we are finding is that regions where the digital penetration is stronger and that producing lower levels of fertility compared to a parallel world where that wouldn't happen. Yes, it's not only a person, right? So because this is social interaction here. So even if you do want to have that kid If all your peers around you are on the phone and they're not paying attention, it's gonna be very hard.
So that's why the digital, going back to the teen story, the digital playground, even if you want to go play baseball at the park, you can go, but it's going to be empty. So it's not only that you want, you need to be off the screen yourself, but you need everybody else to be interacting with you in that. Form that you want. And so in a sense, you don't have a choice now, which is a really interesting proposition for parents, for example, do we take the phone away from their kids? Well, the phone produces isolation. But if I take the porn away, I increase the isolation because everybody else is on the phone.
Juravich: Okay, so yeah, so taking the phones away is not going to suddenly increase our fertility rate unless everyone does it all together at once, right? Were you able to rule out any economic conditions, you know, contraceptive use, child care costs, the cost of housing even, did that contribute to the decline in fertility rates or did you mainly just focus on the technology?
Moscoso Boed: So we mainly focus on technology, but we put everything else, if you will, in a bucket of effects that could be affecting this. So we find that the phone roughly explains 40% of the drop in fertility. The other 60% are coming from this effects that people think about. And going internationally is also interesting because it allows us to say, so for example, we compare the U.S. To England and Wales. And there they have social health care.
And so, and we find almost a mirror image in terms of fertility. So it could be that for some other weird reason, but here we have two countries with very different health care systems that are looking similarly in terms of this connection between springtime and fertility. The same thing goes for the great recession in 2007, the financial crisis. People have said, well, but also 2007 was the year of the financial crises. We do see countries that were not that affected by the financial crises like Australia, and they also show this deep decline in fertility.
So there's a multiplicity of reasons. Each country is talking about their specific forces. As I said, I'm from Argentina and not too long ago, Argentina legalized abortion. And that was the talk. So maybe Argentina's fertility drop because now Argentina has legalized abortion. In the UK, it's about housing prices, but all of them are synchronized right when they get the phone. So it's a very suggestive smoking gun. Not saying that the other effects are not there, but this one is an accelerant of those two.
Juravich: Well, just to end on, now that smartphones are relatively everywhere and everyone has them, will there be a plateau in this drop in fertility rates? Will we figure it out?
Moscoso Boed: That's a really good question, because the one thing that we have now is that the world as a whole is a 2.1 kids per woman. That means replacement rate as global, as humanity. And the question is, how low can we go? Because once we cross, once we're below 2. 1, the next generation is gonna be smaller than the previous one. So how smaller and when is this going to stop? That one, I cannot tell you.
The more extreme cases, so I think the US is around 1.6. So we're relatively below 2.1, but South Korea, for example, is at 0.7. So they are about a third of what they should be to replace the generation. So the drop in their size is really, really fast. Is the world going to look like a South Korea or is it going to more like a U.S. Where we have a more mild decline.
Juravich: I don't know. Yeah, we'll have to wait and see if we can blame the phones. Well, in 10 to 20 years you'll have to do the study again and see if we blame the phone still. All right. We've been talking about iPhones and the drop in the fertility rate with Hernan Moscoso Boed, Associate Professor in Economics at the University of Cincinnati. Thank you so much for joining us today. Thank you. And coming up, we're going to talk about predictive policing. Are you innocent until proving guilty or the other way around? That is when Tech Tuesday from all sides continues on 89.7 WOSU NPR News.
You're listening to All Sides, Tech Tuesday with Amy Juravich, a show where we share stories about science, technology, and the future of our environment. The motto for justice in democratic countries is, innocent until proven guilty. But what if it were the other way around? Instances of algorithms and data analytics being used in policing raise troubling concerns about privacy, autonomy, and freedom. Can police really predict who, what, when, and where of crimes? Or does predictive policing just create a facade of objectivity? Joining us now to talk about predictive policing is a senior writer at Wired, Matt Burgess. Welcome to All Sides, Matt.
Matt Burgess: Thank you for having me.
Juravich: So you analyzed an instance where predictive policing algorithms were being used in Bristol, England. The Think Family database held records and collected data for about half a million people. What was stored in this database? What did they have?
Burgess: So yeah, we were investigating recently this big system of predictive technologies across the region of Bristol in England here. And this database was a kind of collection of highly sensitive data, including mental health records, housing status, teenage pregnancies, and information such as if a child is receiving preschool meals. And this particular database, the ThinkFamily database was Hey! Piece of work that was done by the local police force, Avon and Somerset police, but also the local council, Bristol City Council, which has sort of responsibilities for child safety and schooling and all types of different sort of social welfare issues.
So there was a combination of work done by the police and the council that built this database for one thing. And then as a Building on top of that, they created these predictive risk scoring algorithms that would score people out of 100 in a lot of cases to try and predict the kind of risk that they were going to be and how risky they were going to be. And there's a few different systems that are going on here. So the police have their own, their own set of predictive systems that they were trying to use to predict various different things, such as who may commit a crime on a second occasion.
If somebody was a risk of committing a burglary or if somebody who was also trying to be a victim in terms of a potential crime. And then on the work that the police were doing with the council on the ThinkFamily database, some of those systems were doing the same sort of risk scoring, but also in a way that was trying to maybe identify potential victims of sort of child criminality.
Juravich: So this database was basically like creating a list for the police of people to watch or people to keep an eye on or areas of town to keep and eye on, that kind of thing.
Burgess: Yeah, really, and I think that like there was a few different sort of risk scoring models that were part of this ThinkFamily database, and they sat on top of that data that had been collected, the sensitive data that have been collected and there was two particular risk scoring model that were used in the ThinkFfamily database that were trying to predict instances of child criminal exploitation and child sexual exploitation. In those instances, it was trying to predict which children may be at risk and And then at the same time, the police separately were using some of the same predictive technology to try and identify adults who may be a risk of committing those types of crimes. So it's kind of twofold, really.
Juravich: Did the people of Bristol know that this database existed do they know what information was being stored about them?
Burgess: That's one of the really big things that we found in our reporting that over several years, this database compiling sensitive data and using predictive risk scoring, it had been set up and people didn't necessarily know at all. There wasn't much transparency around this. So our reporting found government reports that were saying that the systems may have got ahead of public awareness of these kind of. Ways that data is being used to risk score people, and overall, like, there are half a million people, around half a million people in the Bristol area, and we think that based on our reporting, not many people would have known about these types of risk scoring systems that were in place.
Juravich: Were you able to figure out if the risk scoring was accurate? You know, I don't know how many years this was, but did people who had a high risk score commit a crime or something like that?
Burgess: Yeah, it's one of the kind of like unknowns about this type of technology in general because there isn't in this case, but in also other cases, there isn't a huge amount of transparency around how some of these technical risk scoring systems are built, so we don't necessarily, in a lot of cases, know what data goes into them, the kind of calculations that the algorithms use to create a risk score, and then also sort of the impacts that are felt down the line. With a lot of the cases that we've seen, not just in this region, but elsewhere, when risk scoring models are used.
The people that are impacted by them don't necessarily know at all that they have been scored. And quite often, there aren't ways for them to find out by trying to access their data or making requests to do so. So I think it's one of the very big challenges that exists with these kinds of technologies is how does it impact people? If it does impact people and the decisions that are made about them, whether that's by law enforcement or by other types of officials, like what is their recourse to challenge these decisions, find out about the decisions that were made. So it's kind of an area that we've seen all around the world, these kinds of risk scoring systems run into problems around accuracy, transparency, and accountability as well.
Juravich: Did the police department and the city council try to justify obtaining and using and doing things with this information? Um, like, or did like, yeah, what was the reaction? Like, I guess from the more, the department level.
Burgess: Yeah, so a lot of these systems are trying to be created to make better use of data that is held by either a police force, or in this case as well, local authorities, and really that they see that they have these huge piles of data about people that they may be using their services or they may have committed a crime if it's the police force and they have years of historic data to try and And like in the age that we're living in, just having this data is maybe not the most useful kind of thing. You have it, you don't necessarily understand all the types of data that you have, but if you can bring it into systems that put it in one place, you can better analyze it using algorithms and predictive technologies.
So you may... Be able to do something that improves your efficiency or improves the value for money that you are providing to citizens in those cases. So I think that there is an aim to try and use data better, and I think on paper it is a very good and potentially noble aim in these cases, but a lot of times the way that these systems are done is often by created, and in this case in particular, there were a small number of people that were using. Or were building these predictive algorithms, they may have then left their jobs in the future.
And then sort of the knowledge of how some of these systems were designed and built in the case that we were reporting, one wasn't necessarily passed down to people who would come in and then be looking at these systems. So in terms of some of systems that were being used in Bristol. We found a independent review into them, and this documentation said that the source code used to run some of these algorithms was not kept. Documentation about how the systems were working and were running was not capped over time. And really I think that if you're going, if you are an authority that is going to be building these kinds of systems, you need to have like good governance and good systems in place to hold. The technology that you're creating to account and just allow people to understand how it's been used.
Juravich: This is Tech Tuesday from all sides on 89.7 WOSU NPR News. We're talking about predictive policing with Matt Bergus, senior writer at Wired. I'm not sure when this whole idea of using this data started, are they using AI to do all of this or does this predate AI being as sophisticated as it is right now?
Burgess: So when we look at predictive policing in general, it's an idea that's kind of been around for at least a decade or maybe longer than that now. And a lot of that is based upon sort of machine learning, which is a subset of AI, but it's basically machine learning is pattern matching. So you have huge amounts of data and you get the machine learning systems to analyze that and find patterns within these reams of data that you have. So within predictive policing systems in general we can have.
Predictive policing systems that will like in this case try to identify individuals and say how much of a risk they are based on sort of characteristics of people that have happened before or been through the system before. There can also be like hotspot based predictive policing. So you might look at police records and see that a certain neighborhood has a huge amount of police reports, incidents, criminology. Incidences of criminality, and then sort of the machine learning algorithms will be able to pick out the sort of like patterns in this that a human may not be able to do if you're looking at a huge spreadsheet or database of files.
And that sort of machine learning approach is something that has existed for for quite a while and is, as I say, it's a type of AI generally, but. It's a little bit different to the sort of large language models and generative AI that we have at the moment, which is really sort of the cutting edge sort of like chat GPT and Google's Gemini and those kind of chatbots. So they're branches of the same sort of technology, but sort of different approaches really.
Juravich: I mean, it hasn't predictive policing kind of been around for a long time, though, even without the computers doing it, because, but I guess it would be called profiling, like police, the police know which neighborhoods, whether they're right or not, that they keep an eye on more that kind of thing.
Burgess: Yeah, they do. And I think that like these kind of technologies over, as police forces in general, have seen the rise of technologies and the use of data, they want to use some of their resources more efficiently. And, I think, that we have seen sort of like predictive policing systems over the last decade. There's been... Some in Florida and Chicago that have been canceled because they didn't turn out to be effective. And I think that at the time, at the moment, you do have a lot of police forces around the world that are trying to use technology to improve their efficiency, to improve the way that they operate.
But you have seen time and time again quite often that these systems kind of fail or are not found to be affective or not found be providing good resources or good value for money. So it is one of these things where the technology kind of does exist to allow some of this to work, but the way that it is being used does not necessarily sort of meet some of the promises that are put out there. And as you say, like we have for a long time, like local police forces know their neighborhoods, they know often individuals who may be at higher risk.
And I think that when you are using technology, it can sort of amplify some of that problems that may exist in data that you. The data that you have. So if you have areas that have been over-policed for decades and then you're trying to sort of like analyze patterns in data there, it's probably unsurprisingly going to tell you that those areas have had more attention and you should keep paying attention to them, which may not be the case in the real world, on the ground. And yeah, it is one of those things where there are big questions about the value and the sort of efficacy of these systems.
Juravich: Yeah, so that feedback loop, I mean, that's what people who are vehemently opposed to predictive policing, they say it targets minorities, it creates a feedback loop because existing police data is inherently biased. You know, some areas are unfairly patrolled more than others. So all the data is in this area that's patrolled. More. So what what did you hear from people who are just opposed to using the data because it's collected unfairly? Have you heard that?
Burgess: There are definitely a lot of people or a lot of people in sort of the academic space and also civil society that do say that any form of predictive policing cannot necessarily work is, as you say, is based on data that may not be accurate, that may include biases already just through this. The structures of society that we already have, and they will amplify these kinds of technologies. So we have seen a lot of pushback in these, to these systems over the years, and there have been, as I say, there have been some cases where actually, after being used for a while, these kinds of technologies are abandoned by police forces or local cities because of the negative impacts that they can have, and also the impacts that it will have on people's privacy, a lack of transparency.
And really, I think that in a lot of cases where we see these kinds of technologies rolled out, there aren't independent audits of these systems. There aren't mechanisms in place to understand how they're being used and the impact that they're having on individuals or on society, or even on officials and police and law enforcement officials themselves because another thing that we do know is a known phenomenon is the issue of automation bias, and if essentially if you do have people that are looking at results of computer systems day in, day out, they may end up trusting them more than they should over time and relying on those kind of results.
So I think that there are a lot of kind of known flaws and instances around these kinds of technologies. And I think as we do see more and more general policing technologies being used, there are constantly issues that are being highlighted, whether it is with facial recognition systems that have been found to have bias. In those, or even more recently, sort of AI-powered number plate recognition systems, which have obviously spread to thousands of cameras across the US.
Juravich: Yeah. Well, these AI systems, they're supposed to be able to make decisions independently. AI is not supposed to biased or racist, but AI makes mistakes. AI hallucinates things. AI is influenced by whoever is feeding it. So do you think that in the future, this is going to pose a threat to people because we as humans don't know what to do with AI?
Burgess: Yeah, there's a lot of unknowns there. And I think that as we sort of like work out that some of these systems may or may not work and there's more examples of this, there needs to be clearer. Guardrails in place for the AI systems, there needs to be accountability for people that are creating them, there need to be transparency about how systems are being used and how they may or may not make decisions about people's lives.
And in a lot of cases, we do see these kinds of technologies being used to the people developing them or using them will say that they produce results that are then fed to a human and a human will make that kind of decision. But I think in the case where we're reporting in England, in Bristol, with some of those predictive algorithm staff that have been looking at the results of these algorithms, said that over time, they didn't trust them. They sort of led to inaccurate results.
And they would know that people that they would expect to be being flagged by these algorithms were not being flag. So I think that the way that these systems are deployed, it does cause and it is already causing real world harms to people and we need to, there are a lot of concerns about AI in general around sort of big existential threats to society, to the world. To life, I guess, in some instances. But I think that at the moment and for years, we have been seeing harms being caused by AI and machine learning systems that people need to be able to, yeah, scrutinize, understand and hold to account really as well.
Juravich: Well, just to end on the U.K. Government is rolling out a $75 million project called Police AI. So do we need to have you back once that's up and running for like a year and find out if it works?
Burgess: Yeah, I think that there's going to be there's gonna be things that there may be some efficiencies, but I think there's gonna be more stories here of things going wrong over time.
Juravich: OK, well, we will talk about it. We've been talking about predictive policing with Matt Bergus, senior writer at Wired. Thank you for joining us today, Matt. Thank you very much. And coming up, the World Cup is testing America's bandwidth, quite literally. And we'll also hear about an 80s-style weather app. That is when Tech Tuesday from all sides continues on 89.7 WOSU NPR News.
You're listening to All Sides, Tech Tuesday with Amy Juravich, a show where we share stories about science, technology and the future of our environment. The World Cup brings people from around the world to cheer on their home teams. Can someone visiting from Argentina video chat with their friends at home in a stadium packed with sixty nine thousand people? Plus, we will talk about how Texas is trying to keep kids safe online. And do we need more AI-powered learning tools? Google says yes. Also, do we need a retro-style version of the Weather Channel app? That answer is also yes. Joining us to discuss these topics and more, we have Russell Holly, Director of Commerce Content at CNET. Welcome back, Russell.
Russell Holly: Thank you.
Juravich: So starting with the World Cup, the World Cups being played all around the US, it brought thousands of people and their phones to the same place at the same time. Does the US literally have the bandwidth to handle this?
Holly: It does. And it's because of an enormous amount of work that carriers have been working on for years now to make sure that not just the stadiums that involving the World Cup, but really, eventually, every stadium in the US is able to handle, you know, a full group of people without losing any functionality on your phone.
Juravich: Yeah, your team ran some tests with different people having different phones on different networks And you tried to send videos send texts all different kinds of things were happening What were the results was one carrier any better than the other?
Holly: So it really depends on the kind of work that you're doing. A lot of the stuff that, there are basically two different services that are using different versions of technology here. AT&T has a thing that they are calling TurboLive, which is actually a one-time pass that you can buy for, that gives you a much higher speed access to a network during things like concerts or events and things like that. So it just dramatically ramps up what it is that your phone is capable of for like a brief set of time in a very specific place.
So, in that test, very little that we were able to throw at that ran into any problems. We're sending videos to someone, it would arrive within 60 seconds. Text messages arrived right away, video chat wasn't a significant problem. This is a huge thing in a space where there can be as many as 70,000 people also using their phones. Verizon had a very similar setup. Instead of having an individual pay system, Verizon is actually working with different stadiums to install additional 5G antennas underneath the seats in a bunch of different places so that you don't have to worry about whether or not you've got four bars, because the bars are basically touching you if you're sitting down for the service that goes into it.
So by setting up what is essentially a mesh network across the entire stadium, You're able to use your phone as though you're sitting at home, which is not something that has been possible before.
Juravich: We have come a long way. I mean, from a time whenever I couldn't get a cell signal at my sister's house to now we're gonna literally sit on our cellular network, wow.
Holly: Yeah, that's absolutely true. You know, here in Baltimore, Maryland was a test bed for when 4G was a very big thing. And I very much remember being in Raven Stadium when they were first turning on 4Gs for everyone to be like, you can actually download a picture, you know, for it being this like massive accomplishment. And that doesn't feel like it was all that long ago, but I am regularly reminded that I'm older than I think I am.
Juravich: Yeah, same, same. If someone had an older phone though, and they're in this stadium, will it not be as good as a newer phone? Did your team test with mostly newer phones?
Holly: We did test with mostly newer phones, but really any phone that has, you know, what is called wideband LTE. So this is, you there was an original version of 5G that came out, you a couple of years ago, and then there was a dramatic expansion of that across Verizon and AT&T and T-Mobile. For this this wider band what that essentially means is if your phone has if you acquired your phone within the last four years There's a very good chance that you know You are a part of this If you go back older than four years then you would need to check the specific model of your phone to see whether that phone released with that ultra wide band functionality But for for the most part if your fun, you know was within the four years Then this is something you get to benefit from
Juravich: This is Tech Tuesday from all sides on 89.7 WOSU NPR News. We were just talking about the World Cup and cell phones and networks. And now we're going to talk about app age restrictions. We are talking with Russell Holley, Director of Commerce, Content for CNET. Texas requires app stores to verify users' ages and get parental consent before downloading apps or making in-app purchases. Is this the solution for keeping kids safer online? Or some are saying it's a violation of free speech.
Holly: Yeah, this is an ongoing conversation and it's something that multiple states are trying to tackle through different ways. California and Texas seem to be kind of leading in very different approaches. California is taking the approach that it should be up to the provider of the service to be able to know whether or not they are serving content to someone who is a minor.
Whereas, Texas is placing the responsibility on the user to demonstrate that they are of age to access certain parts of essentially the internet. And if they can't make that demonstration, then they just don't get access. The downside to that being is you're essentially sending a copy of your government ID to companies like Google and Facebook and you know places where Um, you know, historically data privacy hasn't exactly been a priority, uh, which, you know, puts your personal data at risk in order to demonstrate that you are above 18 to access some of this content.
Juravich: So basically in Texas, the age verification law, to make sure that you are old enough, you have to send a picture of your driver's license or something like that. It's not just, you have just check mark the box and say, I promise I'm not lying, I'm 18. Is that right?
Holly: That's exactly right. Yeah, there is an actual verification process that was implemented as a part of an age verification law that rolled out last year. It was challenged initially in the courts and it had been kind of on hold in the courts. The Supreme Court ruled earlier last week that there was no stay on this age restriction while the court case meters out.
So. For the moment it is the law in Texas that you need to verify that you are above age 18 in order to access a lot of different kinds of things on the internet and if you don't have the ability or you're not interested in providing an image of your driver's license in order to make that verification then you just don't get access without doing some interesting things to bypass those kind of age gates.
Juravich: Yeah. So this is called the Texas App Store Accountability Act. And the U.S. Supreme Court basically said that it can remain in effect while other people are challenging in other courts, which is going to take a while. But like if I if I go to Texas, can I can I get stuff like how does it know who who's who lives where and is doing what?
Holly: So it is entirely geographic. It's the same way that Virginia and several other states now have age restrictions for adult content on the internet where you have to demonstrate your age. This is just a far more restrictive version of that. And if you are accessing the internet via an IP address that is known to be a location within the state of Texas, then you are subject to this policy.
Juravich: Okay. So if I go on vacation to Austin, Texas, and I want to look up something, I'm going to have to take a picture of my driver's license.
Holly: Yes, it is the exact same way that if you are traveling in Virginia and you want to consume adult content, you know, regardless of where you live, it will pop up, you know, an error message telling you that you need to verify yourself before you do anything.
Juravich: And do you have to keep doing it or does it remember you after one time doing it?
Holly: It will remember you after one time with some exceptions. It doesn't happen, it is essentially per device. So if you're looking at something on your phone and you send it over to your laptop and you're not sharing a network, then the laptop will also need to be part of that verification process. It is very much putting the work on behalf of the user in order to do these things.
Juravich: All right, well, we'll have to keep an eye on that to see if it stands after all these court challenges. But now we're gonna move on to Google. Google is rolling out many large language models to help students learn faster. And now the company wants students to upload their textbooks. So Russell, Google has a few AI learning tools. Notebook, LM, Learn About, and Now Learn Your Way. Do we need all of these? And are they all different?
Holly: They are similar concepts aimed at reaching people who learn differently, I guess is the best way that I can explain that. So NotebookLM is something that applies to not just kids, but it's something that anyone can access. And essentially it means the way that Notebook LM works is you give Notebook LM a very specific amount of information and it will only use that information in order to create a chat bot, basically.
So if you feed it a book. You can then ask it questions about the book and it will have consumed that entire book as well as any additional information you provide it and will use that to either create a synopsis or tell you what page something is in or if it's a technical manual. You can say, how do I fix this part here? And it will use the information in the book to provide you a quick answer instead of having to go and look through that book.
Learn Your Way is... Uh, takes that concept and focuses it on textbooks and applies known learning techniques, uh, to help create, uh different environments for you to learn in outside of just reading the book. Uh, so, you know, instead of, if you are a more visual learner, it will, you know it will create images based on the information provided and use that as a, as a storytelling technique to help you, you know, grasp a core concept.
Juravich: Do we still have textbooks? I, is that a silly question? When I saw that, I was like, wait, do we have textbooks.
Holly: This is very much focused on college students. College students who are still very much buying textbooks and in many cases are buying digital versions of textbooks, which has been an option for quite a few colleges over the last couple of years is you're essentially buying an ebook version of your textbook. Taking that information and uploading it to your own personal kind of learn your way account.
Uh means that the information being provided to you is just focused on that book just focused on that textbook so there you run far fewer risks of things like hallucinations uh or you know kind of uh pulling from unverified sources because you have given kind of the restrictions for what information it can build on. And so whether or not this is a good thing is something that I think is going to require a lot of testing from actual educators and not just educators who have been paid very well by Google.
But as a concept, it does take one format of knowledge and shape it in a way that will reach people who uh, you know, uh, maybe struggle with reading in, in, you know, kind of, uh long segments or, you know, find the, the technical manual. If English is not your first language, you know, finding the, the technical manual that you have applied something that can be, you know kind of explained in, in different terms, uh that the potential for it to be useful is certainly there, but whether it is actually something that will benefit, you know, growing minds is something that I think is yet to be proven.
Juravich: Okay. And if once the textbook is uploaded into this, learn your way for Google, does each individual person have to do that? Or does Google have the textbook? Like, is this going to...
Holly: Each individual person needs to do this. The entire concept behind Notebook LM is that it's only the data you've provided it. So it's not grabbing from other sources. So if you have left notes in your book, for example, then those notes are just for you. They're not something that if somebody else were to upload that textbook, they wouldn't have access to that information.
Juravich: Okay, so what you learn from your professor combined with the book can become your own Google Notebook. Exactly. Okay. We only have two minutes left, but I want to have time for the Weather Channel. They've released an app with all of the old school graphics and sounds, and it looks like a lot of fun, but why do we want the retro Weather Channel app?
Holly: This is something that started as a joke for the Weather Channel last year, where they released that there is a specific website you can go to, where it's just a web page that is the original Weather Channel art from the 80s. And the music that went along with that, if you remember sitting in a doctor's office with the Weather Channel music kind of playing on repeat in your brain, this is very much that experience.
And so this now exists in an app because it's everything has to have an app now. So you have the choice if you are sitting at a desktop or a laptop and you want to enjoy the classic weather channel vibes on your app, then you certainly can. Or you can use it on your phone. The important thing here is it is still actually functional. It does still provide useful weather information. But it is done through a lens that I think folks who very much remember that sound from the 80s and 90s will be able to appreciate.
Juravich: That's what I was wondering. So I have the Weather Channel app, so I can go in and I can switch it to be retro, but my weather will still be up to date and modern, right?
Holly: That's right.
Juravich: Okay, and will my radar still work the same, but will it look like the 1980s, but still tell me when it's gonna rain?
Holly: There ends up being some dissonance because radars are quite a bit more evolved than they were when this artwork exists. So you get kind of the retro art and then like a box where a modern radar exists, which is a little jarring, I will admit. But yes, you do still get, you know, kind of modern information through this kind of old school visual. It is a silly thing, but I have to admit, I'm glad it exists.
Juravich: Okay, so all right. So we're happy about it and it and and everyone should give it a go, right?
Holly: Yeah, it's just a silly thing. If you remember that experience, then it'll put a smile on your face. And that's about it.
Juravich: Well, we have been talking about app age restrictions, retro-style weather apps, and more, with Russell Holly, director of commerce content for CNET. Thank you so much for joining us, Russell.
Holly: Thanks for having me.
Juravich: And this has been Tech Tuesday from all sides on 89.7 WOSU NPR News. I'm Amy Juravich. Thanks for joining us.