Plain English with Derek Thompson - Why Are Robocalls So Hard to Stop? (Plus: Kamala and the Gender Wars.)

Episode Date: August 2, 2024

Derek offers his thoughts on Kamala Harris, the new 2024 reality, and gender polarization in the "boys vs. girls" election. Then we talk about the spam apocalypse. The average American receives one spam call or text every single day, adding up to tens of billions of robocalls and texts per year. Derek welcomes Joshua Bercu, the executive director of Industry Traceback Group, to talk about the history and technology behind robocalls and texts, why it’s been so hard to hold robocallers accountable, how spammers do that thing where they make a call look like it’s coming from a local friend, how we've managed to crush certain kinds of robocalls, and what it would take to finally win the war on spam. If you have questions, observations, or ideas for future episodes, email us at [e-mail]. Host: Derek Thompson Guest: Joshua Bercu Producer: Devon Baroldi Learn more about your ad choices. Visit podcastchoices.com/adchoices

Greetings, it's Mal. Call your banners because it's time to head back to Westrose for House of the Dragon, season two. The ringers dragon riders will soar alongside you each week with a heron-hall-sized slate of conversations. The dragon has three heads, and on Sunday nights immediately after Hot D concludes, Chris Ryan, Joanna Robinson and I will be with you for Talk the Thrones. Then on Mondays, two more shows away. Dan Lath and Charles Holmes, Steve Allman, and Jomea Denneron, aka the Midnight Boys, Pugh!

Pee! We'll head to the tourney grounds to share their reactions. And of course, Chris Ryan and Andy Greenwald will sip the Arbor's finest vintage on the watch. Then on Tuesdays, Joanna and I will head to the bowels of a pleasure den for our House of our deep dives. Then on Thursdays, Joe, Neil Miller, and Dave Gonzalez will gather the Ravens for trial by content. In this season, full episodes of Talk to Thrones, House of Ar, and the Midnight Boys will also be available on video on Spotify and the new Ringervverse YouTube channel. Podcast episodes available on Spotify or wherever you get your podcast. Today, we begin with an election update of sorts, and then the main event, today's episode,

about robocalls, robotext, and one man's quest to stop the spam apocalypse. So first, 2024, it's been roughly two weeks since Joe Biden dropped out of the presidential race, and Kamala Harris became the presumptive nominee for the Democratic Party. I think it is fair to say that the rollout is gone just about as well for Harris and Democrats as they could have possibly hoped, Harris is clearly doing better than Joe Biden in national polling. She is clearly doing better than Joe Biden in swing state polls. She has effectively turned what was looking like a clear Trump victory into a toss-up election.

Now, yes, it's a toss-up election with an electoral college that leaned slightly to the right, but that is much better for Democrats in the situation that they had one month ago. Now, right now the media is fixated on a couple issues in the election, like the vice presidential decision. But it's important to remember that just because the media is focusing on something doesn't mean it actually matters. And it's maybe a lesson to carry forward with us

just because the media focuses on something doesn't mean it matters. Vice presidential candidates rarely move elections. Now, they can matter a lot, as we've just seen for a party, if the president steps down or if the president can't run again because he or she has won two consecutive four-year terms.

But right now, a lot of people are trying to scrutinize the difference between, say, Josh Shapiro, the governor of Pennsylvania, a swing state, and Mark Kelly, a senator from Arizona, another swing state, although one that is right now leaning considerably toward Donald Trump. Shapiro and Kelly are not going to swing Pennsylvania or Arizona by five points, or four points, or three points. Two points would be a miracle. one point is exceptional, 0.5 points

would be more in line with the historical average. Vice presidential candidates rarely, even in their own home state, rarely turn elections. So at this point in the race, if you really want to focus on what matters,

I would say focus on three states. Focus on the polls from Pennsylvania, Michigan, and Wisconsin. That's it. If you want to follow the election efficiently, you want to stay abreast of what's going on, but you also need to like spend time with your family.

get drinks with friends, do perform well at work, the efficient way to follow what's going on is to focus on those three states and the polls, the high-quality polls from those three states, Pennsylvania, Michigan, and Wisconsin. With Arizona and Georgia right now leaning toward Trump, that means Kamala has to sweep the Great Lake states to win 270 votes, focus on the Great Lake states. It also means, I should say, stepping back that for the second time in three Donald Trump elections, This will be Donald Trump against a woman. Now, typically, that sort of analysis, guy versus girl,

who be rightly brushed off as reductive, right? You'd say elections are about a lot more than the gender of the candidates, Derek. They're about policies. They're about cohort shifts. They're about economic fundamentals. But in 2024, gender might be the single most important access around which this entire election is spinning. In the last few months, several polls have shown that, for one reason or another,

Trump's resurgence has relied on the rightward shift among young men. That is, young men, men under 30, who have dependably voted for Democrats for the last 30 years, are now shifting historically toward the Republican Party. Last week, the Wall Street Journal reported that its polling shows young men now favor Republican control of Congress and Trump for president. Other polling from Daniel Cox at AEI shows young men shifting right faster than any time this century. Meanwhile, several other polls show women moving left on issues like identifying themselves as pro-choice or calling themselves liberal. While men under 30 have become a Republican bloc, single women under 30 remain perhaps the most dependable democratic bloc.

So what is actually new about the situation? That's a good question. And an acceptable answer might be much less than we think. A majority of men haven't voted for the Democratic presidential candidate since 1976, Jimmy Carter. Meanwhile, women have consistently voted for Democrats since 1980 when Reagan teamed up with a Christian right to take on feminism and abortion. He stormed a victory in 1980, but in the process drew a line in the sand that exists to this day,

with women on one side and men, generally speaking, on the other. This past week, I published an article in the Atlantic on the gender gap and how it might decide the 2024 election. And in a process of thinking through all of these different polls showing that women were shifting left with men shifting right, I had a very interesting conversation with John Sides, a political scientist at Vanderbilt University. And Sides told me that when he looked at the data,

he didn't see the gender gap polarizing as fast as other polls. He said, the gender gap is real. It's just not what most people think it is. Quote, the parties are more polarized by gender attitudes than by gender itself. What the hell does that mean? Well, try a thought experiment on for size. Imagine you're standing on one side of a wall. And on the other side of the wall, there are 100 Americans you cannot see, and you have a job.

Your job is to ask one question, and from the answer of that question, accurately guess how many people on the other side of that wall are a Republican? And the hitch is you can only ask one of the following two questions. Number one, are you a man? Number two, do you think men face meaningful discrimination in America today? Now, the first question is about gender. Are you a man? The second question, do men face discrimination, is about gender attitudes. And a court decides that second question will lead to a much more accurate estimate of party affiliation than the first.

And that's because the parties right now, even though they're polarized, they're not united by gender. After all, tens of millions of women are going to vote for Donald Trump this year, no matter what he says and no matter what he does. But the parties are sharply divided and maybe increasingly divided by their attitudes toward gender and gender roles and the experience of being a man or a woman in America. In other words, it's not men are from Mars and women are from Venus. It's Republican men and women are from Mars and Democratic men and women are from Venus. If I had to sum up the difference between the way the GOP and the Democratic Party think about gender and all the debates that fit into gender, right? Masculinity and femininity and fertility and child rearing and trans politics and abortion rights, if I had to distinguish the way Republicans and Democrats think about all those issues in one sentence, it might be this sentence. The Republican Party sees tradition and gender roles as a guardrail,

and the Democratic Party sees traditional gender roles as a straitjacket. To the right, masculinity is about adhering to traditional rigid expectations. To the left, feminism is about breaking free of those expectations for women. To the JD Vance right, childless women are pathetic because a woman's worth is established in the context of her children. To the left, childless women are making just one of many acceptable private decisions because having children is a fully optional life choice. I don't feel confident telling you I know who Kamala is going to pick for vice president. I don't think I feel confident telling you it's going to matter.

I don't feel confident telling you what's going to happen to the polls in the next week or the next month, the next three months. But I do feel confident. I do feel very confident that this is going to be an enduring theme of the race, a skeleton key for understanding the debates and the discussions behind the 2024 election. Gender is turning into the most important axis of polarization in America. and that's what I'm paying attention to. Now, today's episode, Robo Calls, Robo Texts, the Spam Apocalypse. It's somewhat remarkable that when we talk about the smartphone is the most transformative technology of the century, when my phone rings, when it performs the most normal function of a telephone,

I suspect there's at least a 50% chance that it's just spam. Think about that for a second. the most sophisticated technology in human history, and sometimes it barely functions as an incoming filter of bullshit. Surely you experience something like this, not just on your phone, but in your life. What percent of your physical mail or text messages or basically unwanted junk?

Is it 30 percent? Is it 50 percent? Is it 80 percent? Several reports now suggest that Americans receive up to 50 billion robocalls and 50 billion spam texts every year. And rudimentary back of the envelope math suggests that the typical American adult, therefore gets about one robocall or spam text every single day.

And these calls aren't just pests. They are successful. They fraud and bilk Americans out of billions of dollars a year. And sometimes this junk is surprisingly sophisticated. I've noticed, for example, the last few years, an uptick in scam and spam calls from area codes near northern Virginia, where I was born, or New York City, where I lived, right? As if this new age of spammers has found a way to disguise their calls as possible friends or work acquaintances.

Several times this year, in fact, including yesterday, I've received purported texts from my CEO, the CEO of the Atlantic, where I write. The text will go, hey, Derek, it's Nick Thompson. That's the CEO of the Atlantic's name, by the way, no relation. Hey, Derek, it's Nick Thompson here. I'm at a conference. Can you give me a quick call back? In New York Magazine, Charlotte Cowles, the Cuts Financial Advice columnist, wrote an essay in February

about an elaborate scam ending with her putting $50,000 in a shoebox and handing it over to a stranger who drove away. Now, we could very easily take all of this in a very dystopian direction and predict that the rise of artificial intelligence will make everything worse. As AI voice imitations get more sophisticated, it's not hard to imagine some kooky future where people are, fielding calls from a scammer in an Indian call center and using AI to pass as that call recipient's daughter? You know, hey, dad, I need cash. Please forward me $500, $1,000, $10,000. But before we take this to the darkest possible predictions, it's probably worth checking in with reality, which is rarely so terrible as our most dower interpretations. According to the FTC, complaints about robocalls and live callers have declined by more than 50% in the last three years.

Car warranty robocalls declined from about $1 billion per month in January 2022 to just a few thousand by the end of 2022. Student loan robocalls declined by more than 80% in that year. With traceback techniques, we've been able to snuff out several genres of robocalls quite successfully. And that success holds lessons for taking on the larger spampocalypse. Today's guest is Joshua Birkiew, the executive director of industry traceback group.

Today we talk about the history of the technology behind robocalls and texts, why it's been so hard to hold robocallers accountable, how spammers do that thing where they make a call look like it's coming from a local friend, how we've crushed certain kinds of robocalls, and what it would take to finally win the war against spam. I'm Derek Thompson. This is plain English. Josh Burkew, welcome at the show. Thanks, Derek.

Happy to be on the show. Tell us briefly what you do. So I am the executive director of the U.S. Telecom-led industry traceback group, and that is an effort where we work closely with the industry and with government stakeholders, and we work to find out where illegal calls, including illegal robocalls, are coming from. So before we dive into the history of robocalls and spam and scam that inundates us in this age,

let's cover some basics. What are robocalls? When are they legal? And what makes them illegal? So robocalls, I like to think of them as they are calls made with an artificial pre-recorded voice. That's one of the areas that they're triggered in law. and they are illegal any time to your cell phone without your consent and any time to your cell phone or home phone line when they're telemarking without your consent.

Why are political robocalls legal, at least why do they seem to be legal, whereas some commercial robocalls are illegal? It's been a lot of balancing over the years, but political robocalls, if they're your cell phone, if it's a campaign robocall, they still need your consent, but there is a little more flexibility. when those go to home because of things like, you know, First Amendment considerations. And I want to do a little bit of, you know, definition here before we tell the full story. How do modern robocall outfits work? Like, where are these calls coming from? What kind of organizations are sending them and why?

There's a big variety, but they come from all over. We see, as part of our traceback effort, we have traced illegal calls to entities in the United States, Dominican Republic, India, Pakistan, Cambodia, and really elsewhere. Sometimes when they're illegal telemarking calls, that might be a smaller operation that is trying to sell leads and bombard people with calls to try to sell you some real questionable product or service. The scam calls might come from fraud rings in the United States or elsewhere. Could be a small fraud ring. It could be global organized crime. And in fact, we often see very sophisticated calling infrastructure in India and elsewhere that is defrauding folks here and elsewhere in the world

and then laundering that money through Chinese organized crime. So it can be that sophisticated. I've always wondered what is behind the technology that allows robo calls to mimic the area code of the person picking up the phone? Because I feel like one aspect of spam that's gotten much more sophisticated in the last few years is that I will see a call on my phone. It'll have a 703 area code, I guess full disclosure. I'm from the Northern Virginia area,

so my cell phone number starts with 703. And I've always wondered, how do they know to target my phone with the area code that it's registered to? Is there a name for this kind of technology, and where does it come from? So the name of the tactic is called neighbor spoofing, where they look like they're your neighbor,

and it's gone, In the early days, they might have even spoofed your own number to trick you, and then quickly, people caught on that no one's calling you from your own number. But now it's gone sophisticated enough. I'm your neighbor. I'm in D.C. And so some of the bad calls I get, I'm 202, but I get 240 for nearby Maryland. So it's gone that sophisticated where they'll even do a nearby area code. That's not the exact area code. And the question of why they can do that is like so many of the other issues you cover on plain English is a little bit the rise of internet technology. And, you know, we went from a world where only one person had a phone number, phone line, because you wired,

you know, the house was wired to where you can make phone calls over the internet and plug in whatever phone number you want to use on some of these internet-based calling platforms. Do they work? I mean, I feel like on the one hand, I think I'm pretty good at ignoring robocalls. I might be so good at ignoring robocalls and tech spam that I basically don't use my phone as a phone anymore, period. I just don't trust any number that comes in that I don't have a name associated with. But clearly, if this is an economic activity, even if it's a criminal economic activity that is continuing and growing year after year, it clearly is working. Can you explain to us how robocalls seem to work?

And is there a number we can associate with the scale at which they seem to be profiting? So I'm not sure I have a good number to associate with the scale of how they're profiting. But the way they work, one of the ways they work is certainly when we're talking about the very high-volume robocalls, they're blasting calls to everyone they can

as quickly as they can. They're just playing the odds game. You know, they don't need a big hit rate to have success because they're making billions of calls. Just small percentage of those calls working is success in their book. There are also more sophisticated scams that are more targeted. Those are successful in a different way because those will trick people, young and old, technologically savvy and not. They can be successful. Can you give us an example of the kind of sophisticated technique that is more likely to be successful

in this current age of robocalls and tech? spam. Yeah, absolutely. So one of the ones we've seen a lot recently and is, and look, there's been a lot of work to make it harder to spoof numbers. There's a call authentication framework that was put on the phone network, very complex, so I won't go to detail, but it's gone harder. But there's still some gaps, and one of the things we've seen is live callers pertain to be your bank. And they may even know it is, in fact, your bank. So they may know you're a Bank of America customer when they call you. And they call from the fraud line, what looks like the fraud line. So if you look at the back of your card, it may match. And they say, there is fraud on your

account and you might. And then they put social pressure on them, say, it looks like you did something bad. You've got to stay talking to me or you might have police show up at your door. And they just do a lot of social manipulation, psychological manipulation and try to get someone on. the hook that way. I really want to understand the story of the buildup here because it seems from my own personal experience like these robocalls and tech spam and fraud and scams really increased a lot in the last decade. I wanted to see whether that matched up with the official federal data. And it seems from the records I can find that federal records of robocalls and automated scams seem to peak sometime between 2019 and 2022 and really increased a lot in the years since before.

that. Can you explain to me why the 2010s in particular seem to be this golden edge of robocalls? How can we appropriately tell this story of the rise of the technologies and the economy of robocalls? Yeah, and if you don't mind, I want to wind back the clock a little bit first. And there's really two histories, I think, here. In the late 80s and 90s, people were getting fed up. with telemarketing calls, including robocalls, to their home. Not many people had cell phones just then, and Congress took action. So where you asked me about what laws apply, or what's legal or not legal, one of the primary laws here is the Telephone Consumer Protection Act that was passed in 1991. And that basically required your consent for a robo call, certainly telemarketing robocall.

Congress then later gave the Federal Trade Commission authority to build the do not call list that you and I probably have our numbers on where you can't make a unsolicited telemarking call, whether a robocall or not, to a number on that list. So that works well, I think many would argue with some of the unwanted telemarketing calls and robocalls early on. But then, to your point, we hit the early to mid-2010s. and the problem started to go out of control. And a lot of that, I think, is a technological story.

It is the rise of the internet-based calling platforms really made it very cheap and easy to make calls and make a lot of calls. And it made it cheap and easy for you and me. You may remember when we were younger where we're paying for long-distance calls. No one pays separately for long-distance calls. Now you hardly pay more for an international call

that used to be a dollar. to a minute, right? So that has really been the big enablers. It became really cheap and easy to do this. I'm always interested in stories where technology is both an enabler that makes life more convenient and an enabler of crime or inconveniences that make life worse off. One more detail that I want to add here before we talk about what happened in 2019, 2022, is I'd be really interested if you could give me a sense of where these scammers are based. You mentioned, that the price of long-distance calls has declined tremendously. And also, because this is being done over the internet, there's no particular reason why someone mimicking a call in Northern Virginia

needs to be anywhere near Northern Virginia in order to spoof that robocall to me. So can you give me a sense of where these scamming organizations might be based overseas? Yeah. So what we've seen as part of our... traceback effort is that when we're talking about the illegal telemarketing, a lot of that is actually based here and in Canada where it's trying to get you to buy an auto warranty was a very prolific illegal telemarking campaign at some point. That might even be based here. The scams less frequently we trace to here. And we very frequently trace to India, Pakistan, Dominican and Republic, and it depends a little bit on the scams. But India is a big part of this story.

We, U.S. companies rely on India for their call centers. They have a very advanced call center infrastructure, and some of that is used for scamming. When you say India is responsible for the lion's share of scams, and we're also talking about spam, SPAM, and robocalls, are you saying that the operations out of India tend to be banks of people making live calls, that is human voices talking on the phone to people? Or is that also where the robocalls are coming from? It's all the above. So some of the early robocalls would have press 1 to hear more or something like that.

And when you hit 1, you would be connected with a live operator. So even if it was a robocall to really enact the scam, They want to get you talking to a human, right? So they send out the robocall. That's their lead. They get a lead in, and then they start going to work. We have seen a shift from high-volume robocalls as it's been harder to get those and keep those on the phone network to the first call being a live human. We've seen a lot more of that today.

And just out of curiosity, why India? I just, I think it's, there's some obvious reasons where it's a, um, a poorer country that speaks the same language as we do in the United States. And so there's a little bit of that. But also, because of that, we have deployed a lot of call center infrastructure that our companies legitimately use in India. And so that's part of it too, where there are a lot of people trained to work in call centers in India. Most of them legitimate, sometimes they end up at an illegitimate one that scams. That's a big part of it. And it's interesting that I was having a conversation with a company that was based out of Germany and saw something very

different. A lot of their scams were coming from Turkey and sort of the same issue there because there's a lot of German speakers in Turkey. And so that enabled that the scams to come from there rather than what we see, which is often India. Interesting. So you're saying that maybe American companies outsource their call center operations to India. And then people that worked in those legal call centers may have gathered the skill in order to open extra legal or illegal call centers. I mean, you're not making an explicit accusation, but you're saying that might be one causal mechanism that ends with a disproportionate share of these robocall centers being located in India. Is that the right recapitulation? That could be part of it. It also is that because of that,

there is an expectation of a lot of voice traffic that would come from India and end up in the United States because there is a lot of legal legitimate traffic that happens that way. And then the other thing is because they have that infrastructure there, there's a lot of call center jobs, too. So there's a lot of reports and documentaries about this, but some of the people who get started in a legal call center thought it was a legitimate call center, which is a very common job there. And then they realize later that they actually are selling not real Microsoft tech support, but they're actually scamming people, but took them a little while

and realized that they were part of a scam. So I want to catch people up with the narrative so far. There's a deluge of robocalls and scams in the 1980s, 1990s. The U.S. government passes several laws to create, for example, a do not call list in order to shut down this robocall revolution on the previous generation of telephonic technology. But then in the 2010s, you have the rise of the smartphone,

Smartphone penetration screams past 50%. Lots more people are on their phones. The cost of long-distance calls declines. Calling becomes digitized. It therefore becomes easier to make this new generation of robocalls go viral. Robocalls are growing from 2015 and 16 through 2017, 18. And by the spring of 2019, the U.S. government is receiving more than 500,000 complaints a month. This is according to FTC data, more than 500,000 complaints a month of either automated robocalls or voice scams. In the next few years, there is a concerted effort to fight back against the robocall apocalypse. Why don't you pick up the story here? What did we do to try to defeat this hydro monster?

Yeah, we did a lot. And in fact, I think it's a real. really, it is a silver line in the story and really a way of how government and the industry can work together and solve problems. And that's really what we did is there was a lot of work on the industry side. Sometimes government creating the framework for doing blocking and labeling. You probably seem spam likely on your phone when you get that on the mobile phone. All that's been deployed since then. As I mentioned, there's a framework that makes it harder to spoof. And then one of the things is our traceback effort that we started to roll out then.

And nowadays, virtually all, there's more enforcement than ever before. And it's coming from the Federal Communications Commission, the Federal Trade Commission, and the State Attorney General are more active bringing enforcement than they've ever been before. And all that's based on the data that we're able to get in traceback and identify who's making these calls and who's enabling, who's bringing the calls into the country as well. Can you take us inside traceback? You know, I have no expertise in telephone technology. I have no expertise here, but I would love to understand this story at a more granular level

because we talked about this in previous episodes. I think the media is very, very good at writing about, podcasting about, doing television on catastrophe. We're great at shotting the spotlight disasters. So when the number of robocalls and FTC complaints screams past 50,000 a month and sets a new all-time record, we're very good at saying, oh, robocalls have never been worse. This is the worst it's ever gotten. It's only going to get worse in the near future. We're less good, I think, at reporting on when problems are solved and how problems are

solved. And in fact, in March 2024, the last month with complete records that I could find at the FTC website, the government received 220,000 complaints. Now, that's still a lot of thousands of complaints about robocalls and scams. But that's a 50% improvement over just five years. And it seems like this traceback operation was a huge part of our success in reducing the number of robocalls. So just take us a little bit into this at a level that we can understand. How does this actually work? Yeah. And I think you raise a good point that while we've made tremendous progress, there's still a lot of complaints. And so that's one of the things we face, too, where it's actually similar to allow your coverage of the economy,

that the indicators are actually showing strong economy, but it's still hard for people to buy their house. It's still expensive to go to a restaurant. And, you know, robocalls are like that. I don't get many today. My father-in-law tells me he gets 10 a day. So we do have that too where not everyone has the same experience. But just in terms of traceback,

the phone network is not that different from when you think about shipping. So if I order something online, you know, with the wonder of Amazon, it's at my house in two days. But that wasn't that product's full trip, right? It might have made overseas, got onto a boat, you know, one carrier took it across the world. It got to a train. It got to the last truck, et cetera. The phone network actually works somewhat similarly. if a call coming from, let's say, India or the UAE or China, it doesn't necessarily go straight

from the phone provider that put the phone call on the network to me, if it's coming to me. There's actually a series of different relationships and providers that all will touch that call instantaneously, but all will touch that call to get it to be delivered to me. And that's what we do in traceback is, you know, you talked about spoofing. If you can't trust the number, there's no way to just look at the number and say, here's where that call came from. So that's what our traceback effort does, is we walk back that and we start with the, if that call hit me, my provider, my voice service provider, and we go to them and say, where'd you get that call? We just walk back that line until we find out exactly where the call came from.

Is there a particular reason why we've had more success with certain types of robocalls? According to the data that your team shared with me, car warranty robocalls declined from about one billion per month, one billion per month in January 2022 to just a few thousand by the end of 2022. student loan robocalls declined from 35 million in January 2022 to 5 million in December. So with car warranty and student loan robocalls, I mean, talk about success stories. You have a decline here of between 85 and 99 percent decline in spam. Why so much success with these particular categories of spam? You know, I think one reason was they were really prolific. And that got our attention and it got the attention of the regulators. And so we very actively traced those campaigns and the Federal Communications Commission and the states took very decisive action. And basically that set the standard that for folks in the industry, smaller voice service platforms that when they get When they see someone saying, hey, I've got an auto warranty business, I want to put my calls on

your network, they knew that that was probably suspicious because I've been made very clear by the regulators based on our data. If I wanted just a little bit more detail in my visualization of how traceback works, I love the way you compared it to a supply chain. When I go to the FedEx website or the UPS website, and it's showing me how some item, you know, It's a book. It's a charger for my cell phone. It'll say, you know, this item, the order originated in Tennessee.

And then it was sent to a shipping center in Maryland. And then it was sent to a shipping center in Northern North Carolina. And now it's with a courier that's going to come outside my door in Chapel Hill. So there I can see very literally, oh, if I was going to be a detective to reverse engineer the supply chain, I know exactly where to go. I need to drive north and North Carolina, that I need to fly, you know, to Tennessee. Can you just give me a little bit of a clearer sense of how this equivalent operation works for traceback? Yeah. So how it works is we get an example or many examples of an illegal call campaign or a legal call.

And we get those a few different ways. It may be you mentioned the complaints to the FTC or complaints to other federal. and state agencies, and they'd refer them to us and say, hey, there's this call going. We've actually sourced working with different partners, a lot of data on our own where we get examples of auto warranty. We were getting from different partners hundreds and hundreds of examples per day of actual calls of that that we had from honeypots and other entities. Or we've been working with financial institutions, the same thing. They'll refer some over to us. And that's where we start and we have that actual call example and we go to that carrier

and we say, who gave you this call? Because I think when you talk about the supply chain, it's similar in that what each carrier knows who touched the call, all they know for sure is who gave them the call. They don't necessarily know where it started. Usually they don't. They just know who gave them the call. So that's how our system works as we go to them and we go, who gave you the call? Now, it's not manual. It's not us calling them up. That's how we started. We started this with emails in a spreadsheet, but we built a lot of technology along the way. And so we have a portal that the providers will log into and it's automated where they'll give us those responses automatically. Sometimes when we make progress in one channel, the channeling just moves somewhere else.

It's like squeezing a balloon in the air or the water just flows to some other part of the balloon. And the place where another part of the balloon has inflated with your success in stopping robocalls is that a lot of the spam and scam is moved to text messages. I feel like in particular over the last three years, I've gone from getting almost no spam text messages to about every other day I get, hey, or, hey, what about the dinner that we were supposed to go to? Sometimes they're incredibly familiar. And I can sometimes just look at this text message from a number that I don't have my phone and think, wait, did I accidentally make a dinner reservation with someone in North Carolina and not collect their number? Can you give us a sense of when and how the locus of spam activity migrated from robocalls to robot texts? Yeah, I mean, I think it aligns pretty close.

with some of the data you cited. As we saw robocalls declining, we saw that the scam calls or scam tax going right up. And I think you're right. It's related. And it's because it got harder and harder to make robocalls. And we have to remember whether it's the spammers or the scammers. This is their livelihood, right? This is how they, this is how they defraud people or get you to pay for things you don't want to put food on their table. So they don't give up. They don't just go, okay, this was fun, I'm going to go get a different job. They look for a new channel and a new way to efficiently carry out their schemes. And so that's exactly what we saw.

Is it the same technology, essentially? I mean, it's just, it's the same automation of words just typed rather than spoken. So are we looking at essentially the same technology? And if it's the same technology, I guess the follow-up question would be, why has it proven so hard to have the same success against text messages that we seem to have had against robocalls? So it is not the same technology. There's a lot of distinctions, including that in the text world, it doesn't quite work. The supply chain isn't quite as distributed, as complex as it is in the call world. So there are some distinctions.

In terms of success, I think there's a lot of work being done, and it's sort of a newer problem. With the robocalls, it took us some time to start having the success we did. And we did then have that success. And I'd say, you know, we're probably at that end of the curve on the text messages where there's reason to hope, but it's still early on deploying solutions. Last quick question for return to the future in artificial intelligence. is it your expectation that the same kind of traceback technology could also prove useful with text messages? Or am I right in interpreting what you just said as we need an entirely new different kind of playbook to have the same kind of success with spam texts that we have had with spam calls?

So I think it's always worth looking at what worked and didn't work and seeing what works and seeing if it can be applied to a new use case like text. But I do think there's some distinctions that it does require some different solutions overall. Let's talk about AI. There's a lot of fear right now that AI is going to bring about a golden age of spam and scam. We hear news reports about the power of so-called deep fakes, news reports about AI being able to imitate the voices of intimate, So, you know, you imagine some dystopian future where vulnerable grandmothers and grandfathers across the country are picking up calls from numbers that maybe look like their grandchildren,

and the voice sounds just like their grandchild, and it's telling them that they're stranded on the side of the road somewhere when they went on some college trip, and can they please wire $5,000 to X and X account? I think Warren Buffett, you know, when he was shown his own image and his own voice, just very convincingly replicated by an AI tool, essentially said that if he was interested in investing and scamming, it would probably be the growth industry of all time. Give us a sense as an industry expert of what risks you think are overblown when we talk about the intersection of AI and spams, and what risks you think are reasonable that we should

prepare to look out for? So I think the risks are real. And I think there's good reason to be concerned. In particular, to that point, I think Genre of AI can help improve quality of scams. Now, I don't think in the near term, we're going to see every scam pretending to be your grandchild to be using AI, because honestly, they're having enough success without it. So that's a sad part of the story is that they already are, it isn't worth the extra time and investment for some of the scammers right now because they're able to defraud people so rapidly. But in turn, one of the fears I hear about Generv AI is, well, not only will improve the quality, which I think is a real risk, the quality of scams, it will also

make it easier for, I mentioned India. We get a lot of our scams from India. But if you can rapidly change languages and have a conversation where Generva.I. is doing it. Well, now some of those folks who only speak English will also be defrauding Europeans. And meanwhile, we might be gained some fraudsters from some of the places we haven't seen them historically targeting us. Those are all real risks. Some that I think are overblown is the notion that every call will be AI. The tools we built into the network to help solve for the robo-call problem, they're really good at seeing someone making thousands and thousands, hundreds of thousands, millions of calls. So if that's the fear, hey, we can scale up these sophisticated attacks and do thousands and thousands

per day. That's something we could see. What's a lot harder is when they make five really bad calls. The other thing I'd say is, you know, we already are tracing back some of these calls. So traceback works the same, whether it's made with AI or not with AI. And we in January for the Democratic primary when there was the Biden robocall, New Hampshire that was pertaining to be the president telling people not to vote, we did successfully trace that back, actually finding the source within just a couple hours of when we initiate tracebacks, and that led to a lot of rapid enforcement after that. So traceback is fighting the supply side of the spam apocalypse.

I also think it's important to point out that there are behavioral changes happening on the demand side. When I reflect on my own behavior, my changing attitude toward my phone, as I said, I just don't pick up my phone anymore. I just assume that if it's a number that doesn't have a name already attached to it, right? If it's not Josh calling me, it's someone from Josh's area code, I'm just going to let it go to voicemail. And if the robocall wants to leave a message, well, guess what? I don't listen to my voice messages either. So it's just screaming into an ether. And it's possible that, you know, I'm 38 years old, but there's people who are 78, 88 years old, who are receiving calls. As you have this generational replacement effect where people like me get

older who just don't pick up their phone or respond to text unless they have 100% certainty about who's calling on the other end, that maybe changes on the demand side will help to resolve the robocall apocalypse on its own. Do you guys think about that? Is there any evidence of it that you sort of see in the data that suggests that people are just picking up their phones less? Yeah, I think people are picking up their phones less, but, you know, I think our hope and a lot of our work is centered around how can you trust your phone again, right? And how can we build both protections that a lot of the carriers have deployed, but also with things like traceback accountability on the people creating the problem in the first instance and hold them accountable

so that, you know, if you want to, you could pick up the phone and trust that it's your, you know, your dentist and not have to always let it go to voicemail. That's our goal in the end of the day, is how do we get back to there? So this is not mission accomplished. There are still hundreds of thousands of complaints a month. You just said a few minutes ago that one reason why spammers might not necessarily invest too highly in generative AI is that they have enough success with this generation of technology. But I want to close on a note of optimism that the fact that robocalls, or at least robocall complaints to the FTC, have declined by 50% in the last five years, or more than 50% in the last five years, is absolutely a policy success story. In a way,

this is a perfect model of exactly how we want government to work. There was a problem in the real world. The government collected data to illustrate the scale and growth of this problem. We had agencies work with the private sector to resolve this problem, and then we had progress on this problem, demonstrated quantitative progress on this problem. If you were going to think of traceback, And I'm not trying to inflate you too much, but maybe that's what's happening. If you're going to think of the traceback program as a kind of model for how government and government-adjacent groups can be responsive to public problems, what is that formula? I think the formula is that. It's the partnership. And it's an area where the private sector actually went to the government and said, hey, some law and regulation here will be helpful because we're already doing this. but we need the whole industry to be doing a lot of this because it is a network.

You know, when we talk about any issue on the internet, it's all interconnected. And so you need to push accountability throughout. So I think that's part of it. And then the other thing I would offer is just in terms of the robocall problem and continuing to make progress and optimism is, you know, I know you track the NBA. And in particular, I think I've heard you talk about it, and I'm similar where I, probably track NBA gossip and train machinations more than I actually watch games these days. Guilty is charged.

Yeah, right. So, you know, while it didn't necessarily work for the 76ers, I do think this is an area where you say you need to trust the process because it has been working. The data shows it. And some, it's not always satisfying to policymakers, but I think it's an important thing to remember, which is incremental. progress is progress. And sometimes we have to just keep making progress because there aren't magic

wands to these solutions. There's no piece of law you can write in tomorrow and these problems go away. And so that's the big takeaway, I think, for hard to solve problems. Josh Berkew, thank you very much. Thank you. Thank you for listening. Today's episode was produced by Devin Beraldi. Our summer schedule for plain English for the next few weeks will be one episode a week on Fridays. We'll see you next week. Thank you.