Delia Paunescu | Vox Our world has too much noise and too little context. Vox helps you understand what matters. 2019-12-06T18:19:48+00:00 https://www.vox.com/authors/delia-paunescu/rss https://platform.vox.com/wp-content/uploads/sites/2/2024/08/vox_logo_rss_light_mode.png?w=150&h=100&crop=1 Delia Paunescu <![CDATA[The faulty technology behind ankle monitors]]> https://www.vox.com/recode/2019/12/1/20986262/ankle-monitor-technology-reset-podcast 2019-12-01T10:42:14-05:00 2019-12-01T10:50:00-05:00
An ankle monitor is seen on a migrant woman from El Salvador. She was released from federal detention with fellow Central American asylum seekers at a bus depot in McAllen, Texas, on June 12, 2019. | LOREN ELLIOTT/AFP via Getty Images

Electronic GPS monitoring within the criminal justice system isn’t widespread. But it’s become more prevalent in recent years. 

In 2005, around 53,000 people were supervised with monitors, according to the PEW Charitable Trusts. By 2015, that number had reached more than 125,000 people. That’s a 136 percent increase in just 10 years.

Some people see the rise in electronic monitors as a positive alternative to mass incarceration because it lets people pay their debt to society while still providing for their families, host Arielle Duhaime-Ross explains.  

But as this episode of the Reset podcast uncovers, the technology of ankle monitors doesn’t actually work very well, which means it ends up having hugely negative impacts on the lives of the people it’s meant to be helping.

Sarah Faye Hanna, a 34-year-old mom from Pipe Creek, Texas, wears an ankle monitor as part of her three-year probation sentence. She was convicted of drug possession in 2018.

Hanna says her ankle monitor is about the size of a packet of cigarettes. It lights up green when her location is available and red when it isn’t — which happens more often than she’d like. When the ankle monitor loses its GPS signal; she has to walk around outside until the signal returns. She also has to spend long periods tethered to a wall while the ankle monitor charges, which can make taking care of her newborn tricky. Hanna pays $300 a month to have it.

“I just I hate it. It embarrasses me … I feel like it’s a little bit extreme. They don’t have justification to do this. I’ve gone to court. I’ve done everything that I’m supposed to do. It’s a lot of hassle and it’s super expensive. It’s an invasion of my privacy,” Hanna says.

There are many different reasons why people are ordered to wear ankle monitors. One is probation, as in Hanna’s case. People on parole may also have to wear them, after they have served their sentence. The US federal government uses them to monitor undocumented immigrants, as in the photo above. And people awaiting trial sometimes have to wear them, even though they don’t have a conviction.

Shubha Bala, director of technology at the Center for Court Innovation, describes a 2015 study she conducted with the New York County District Attorney’s office in which young people got the ankle monitors as a possible alternative to spending time in jail at Rikers Island. She observed that the monitors, which connected to the defendants’ phones, are often disruptive. In one instance, a student was kicked out of class because his phone had detected an issue with the monitor and it wouldn’t stop beeping.

“Would I personally feel like electronic monitoring was a good way to continue? I would say no. Some people see electronic monitoring as this magical way of knowing where someone is all the time. And that’s not what it is,” Bala concluded.

Later in the episode, Duhaime-Ross has an in-depth discussion with Robert Gable, the man who —along with his brother — has been credited with conducting the first experiment on electronic monitoring in the 1960s, while the pair was studying at Harvard University.

Gable and his twin brother Ralph set out to see if they could stop young people from committing crimes through a number of interventions, including by monitoring their movements and giving them rewards for good behavior.

Listen to the entire conversation here. Below, we’ve also shared a lightly edited transcript of Bala’s conversation with Duhaime-Ross.

Listen and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Shubha Bala

I’ll often hear, “But isn’t electronic monitoring better than jail or better than prison?” That’s the wrong question. My really facetious example is: Giving people ponies is also better than being in jail. But we don’t do that because it’s just a useless alternative.

Arielle Duhaime-Ross

In 2015, Shubha studied the use of electronic monitoring within the criminal justice system in New York City. The District Attorney’s office of New York County was running a pilot program to test the use of ankle monitors on young people waiting for trial — people who might otherwise have had to spend time in New York City’s most infamous jail, Rikers Island.  

The type of ankle monitor used in the program works by connecting to the wearer’s phone via Bluetooth. 

Shubha Bala

We had one kid actually get kicked out of class because their phone kept going off. They couldn’t figure out why it was going off and the teacher said it was disruptive.

In New York, kids aren’t supposed to have phones in class at all. So what it meant is that we were already violating the kid’s right to privacy around these charges because we had to tell the teachers that they’d need to be able to keep their phone in class.

And then you think of what are all the other kids in class thinking. Because this one kid gets to keep their phone in class when none of the other kids do. 

Arielle Duhaime-Ross

Now it’s beeping and disrupting classes. And this young adult, who is presumably being monitored in order to keep them in class, is now being kicked out of class because of this surveillance system. 

Shubha Bala

There is a lot of nuance that doesn’t really get thought of. For [the purposes of our study], it rendered the technology ineffective because we had to make so many exceptions.

This was one example but there were so many times where we had to turn off any alerts during certain hours. Then what’s the point? Why bother having a monitor?

We ended up doing exactly what we did prior to electronic monitoring: having a relationship with the principals of the schools and getting the person’s attendance once a week.

Arielle Duhaime-Ross

The most old-school way of checking in is actually just saying, “Hey, principal, is this teen in class?”

Shubha Bala

Exactly. 

Arielle Duhaime-Ross

So the same technology that was intended to make sure this student was in class became the reason he got kicked out. This student’s phone was going off in class because the ankle monitor was running out of battery and his phone was alerting him to charge it.

But his wasn’t the only problem the researchers encountered. 

Shubha Bala

It only tracks people when it’s on or near the person and it has batteries and there’s GPS signal.

So that’s issue No. 1: it’s not a magic bullet that tracks where somebody is 24/7, constantly, no matter what.

Issue No. 2: It causes a lot of alerts. There’s a lot of information because there are a lot of components that need to be working for the monitor to work. Which means it’s information overload both for the person on the monitor and the person who’s supervising them.

Arielle Duhaime-Ross

It sounds like this pilot project that you ran in New York City was not super successful.  

Shubha Bala

As a technologist and policy person, I would say it was really successful in that we learned a lot of important information. 

But in terms of would we continue? Would I personally feel like electronic monitoring was a good way to continue? I would say no.

Some people see electronic monitoring as this magical way of knowing where someone is all the time. And that’s not what it is.  

Arielle Duhaime-Ross

Despite all these problems, electronic monitors are currently used by the federal government, all 50 states, and the District of Columbia. But if the devices are so problematic, why are they being used so widely? And who’s profiting from that?

Shubha Bala

The majority of big companies that are in this space are companies that have been working in the prison industry for a long time. Those are the companies that are largely expanding their portfolio of whatever else they were offering to prisons and jails. And they’re expanding it by including electronic monitoring as an alternative. 

Arielle Duhaime-Ross

It seems counter to their business to suddenly be working on electronic monitoring when they were previously operating prisons or implementing technologies within prisons. Why pivot to electronic monitoring?

Shubha Bala

It makes sense. Hopefully we’re working on reducing the number of people that are in prison or jail, and that obviously really eats into your revenue if you’re running a prison or jail. This is a really great way to continue to grow and expand revenue even while people are not in jail or prison. 

The decision makers — who I think are well-intentioned people — are not getting the actual nuance, complexity, and difficulty [monitors cause] for everybody.

What’s interesting about criminal justice technology is that the people who are using the technology, both those in the criminal justice system and a lot of the supervising officers, don’t get a voice in the technology. 

Arielle Duhaime-Ross

After I spoke with Shubha, I had a question: How did we even get here? Who came up with this idea? 

I actually talked to Robert Gable, a retired professor of psychology at Claremont Graduate University and one of the people who’s been credited with inventing the use of electronic monitoring in the 1960s. He has some complicated feelings about how the tech is used today. 

To hear Duhaime-Ross’s interview with Gable, listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

]]> Delia Paunescu <![CDATA[What to know before posting a photo of your kids on social media]]> https://www.vox.com/recode/2019/11/26/20983980/posting-kids-photos-online-privacy-sharenting-reset-podcast 2019-12-06T13:19:48-05:00 2019-11-26T15:40:00-05:00
Studies suggest that kids will have well over 1,000 pictures of them on social media before they’re 5 years old. | Photo Illustration by Thomas Trutschel/Photothek via Getty Images

The internet isn’t always the safest place for adults, especially considering the amount of privacy we forfeit every time we share information about ourselves online. But what about kids and babies? How safe is it for them?

This episode of the Reset podcast explores what’s at risk for children when the adults in charge choose to “sharent,” or put out photos and videos about them on social media. 

In the first part of the episode, guests Kara Swisher and Scott Galloway — co-hosts of the Recode podcast Pivot and parents — admit they both post photos of their children online with varying degrees of privacy. “I’m a terrible sharent. I don’t know how else to put it,” Swisher tells host Arielle Duhaime-Ross, saying she’s shared pictures of her children “for a long time.” Galloway, on the other hand, made his Instagram private because he finds it “somewhat performative when other people post pictures of their kids before the age of consent.”

Later in the episode, Leah Plunkett, author of Sharenthood: Why we should think before we talk about our kids online, explains the potential dangers that lurk with having anything about your children on the internet, whether it’s shared to a private account or not.

“We know that a number of the images that are pornographic of children are pictures of real kids that are taken offline and photoshopped or otherwise retouched. We know that children’s identities are particularly vulnerable to identity theft because most children would have no legitimate reason to have their Social Security number tied to a credit product.”

According to her, privacy is a myth when you consider that someone can always take screenshots of your private content or access it if someone who follows your locked account has their phone stolen.

“Any time a picture is shared digitally on social media, you should assume that it could get out of your control,” she warns.

She adds that “any device in your home that is taking in digital data and sharing it back to a software application” poses a potential danger. That means home cameras like Nest, home assistants like Alexa, or the Echo Dot Kids.

Listen to the entire conversation here. Below, we’ve also shared a lightly edited transcript of Plunkett’s conversation with Duhaime-Ross.

Listen and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Leah Plunkett

I think that the desire to connect with people we know and make connections with new people by bonding over our kids is really natural and beautiful. But the technology tricks us into not realizing just how public it is.

Some studies suggest that kids will have well over 1,000 pictures of them on social media well before they’re 5 years old. And those pictures and the information that comes with them, they’re not private.

So even if your social media settings are set to private, all it takes is for somebody to take a screenshot of that picture and then repost it in a public-facing way or for somebody to have crept into your social circle under false pretenses.

So you think you’re sharing with a closed community and you’re really not. I really do worry about thinking we’re keeping something private when actually we’re broadcasting it to thousands, if not tens of thousands, of people.

Arielle Duhaime-Ross

What exactly are the risks? Worst-case scenario, what are we talking about when we talk about people sharing photos of their children on Twitter, Instagram, or Facebook?

Leah Plunkett

So we know that a number of the images that are pornographic of children are pictures of real kids that are taken offline and photoshopped or otherwise retouched. We know that children’s identities are particularly vulnerable to identity theft because most children would have no legitimate reason to have their Social Security number tied to a credit product.

So if you are an identity thief or somebody intent on obtaining credit in someone else’s name, you can get a Social Security number through a data breach like Anthem or some of the other big high-profile data breaches that have had Socials attached to them. And then you combine that with information about age, location, exact date of birth and other details that we parents share readily.

It’s so easy to create false credit applications in kids’ names 

Arielle Duhaime-Ross

And the government doesn’t stop that? Nobody is going, “Hey, this kid is asking for a line of credit. That is weird. This child is 6 months old”?

Leah Plunkett

The government will come after it and bring an enforcement action or a criminal prosecution once it happens. 

Arielle Duhaime-Ross

My main reaction is, 1) people don’t like being told not to share pictures of their kids online. And 2) even if you are the kind of parent who’s going to be really, really careful, there are no systems set up for you to be absolutely certain that you are doing the best by your child.

Leah Plunkett

You are right on both counts.

There is no systems set up. Even those of us who are really trying to think about privacy when it comes to social media are put in a really tough position because short of something happening to our child’s image that violates criminal law — [like when] somebody takes it off the internet and manufactures into it — there are not a lot of law enforcement or regulatory mechanisms keeping an eye on the Wild West of social media when it comes to children’s pictures and their data.

As a parent, you’re really left to your own devices. Do you want to try to parse through all the terms of use and privacy policies for Facebook, Instagram, Twitter, or whatever the next thing will be?

The studies show that most of us don’t read these terms of use and privacy policies. And even if we do read them, good luck understanding them. I say this as somebody who studies them: I have trouble. 

Arielle Duhaime-Ross

What would you say to a parent who doesn’t think that a picture of their child naked in a sink or a bathtub is a problem if it’s shared on a private Instagram that has very few followers? 

Leah Plunkett

I would say that they’re right, that they are taking more precautions than, let’s say, posting it on Twitter and tagging it “#nakedbabyinabathtub.”

But any time a picture is shared digitally on social media, you should assume that it could get out of your control.

It could get out of your control if one of your followers loses their phone and someone else gets it, if one of your followers takes a screenshot of it and re-shares it or has their phone hacked into, and even if you feel really confident that your five Instagram followers are your best friend, your parents, and your grandma — so you trust them completely and feel like the odds of them losing their device are slim — you still can’t control or even really know what may be happening behind the tech company scenes with those images.

So I’d say to that parent that what they’re doing is lower risk than certain behaviors. But it’s not no-risk. My rule of thumb to parents is don’t post pictures in any stages of undress.

Arielle Duhaime-Ross

How is that any different than the risks for adults? Why focus specifically on kids?

Leah Plunkett

Because parents especially (but also teachers, grandparents, and other trusted adults), we have to step in and serve as the legal decision-makers for our kids when it comes to their images and their information.

Any time that you are being asked to step into someone else’s shoes and serve as their watchdog, you have a heightened moral responsibility to make sure that you’re minimizing the risk that they could be exposed to. 

Arielle Duhaime-Ross

I’m not sharing stuff on social media, what other data might be accessible about my kid?

Leah Plunkett

If you are using any smart devices in your home or any sensor-based devices, then they are going to pick up information about your children. So a Nest cam, a smart TV, or a smart diaper (which is a real thing).

Any device in your home that is taking in digital data and sharing it back to a software application or other type of digital infrastructure is taking your kid’s private information and having it leave your home.

Arielle Duhaime-Ross

That’s actually kind of terrifying. What’s the solution here? Are you advocating for parents to become technophobes or luddites?

So much about having a child these days is involved in setting them up for success by making sure that they have early interactions with tech. But it sounds like you’re saying to hold off on that. 

Leah Plunkett

I am advocating that all of us start to just have more conversations about it and raise our own awareness. Parents should look for low-tech or no-tech ways to do particularly intimate things. So, for instance, going back to the smart diaper, if [there isn’t a medical reason for it], maybe just think about changing that diaper the old-fashioned way.

While I’m not advocating for parents to become technophobes, I am advocating for us to make values-based decisions and think when we are using a digital technology or service, whether it is social media to stay connected or an Alexa or other home assistant: Are the benefits from that outweighing the potential privacy risks and the potential downstream risks to children’s current and future opportunities?

It’s very reasonable for parents to say, “My social media settings are set to private so I feel pretty confident that the images aren’t being re-shared. I don’t really care if surveillance technology uses my kid’s pictures and it is worth it to me to stay connected to my family in Australia,” for instance.

I don’t advocate parents break their phones or go live in a hut in the woods. But I do advocate that we all raise our awareness that there are hidden costs and hidden risks to doing things digitally. 

Listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

]]> Delia Paunescu <![CDATA[Your Amazon package has a wild journey before it gets to your door]]> https://www.vox.com/recode/2019/11/24/20978484/amazon-target-walmart-shipping-package-delivery-roundup-montana-reset-podcast 2019-11-24T11:14:56-05:00 2019-11-24T11:15:12-05:00
In a recent three-month quarter, Amazon generated $11 billion of revenue from it third-party services.  | Photo by Europa Press News/Europa Press via Getty Images

As increasingly more shopping is done online, the solutions online retailers devise to meet delivery demands have become ever more patchworked.

One such workaround is something called a prepping center, where packages are delivered from third-party retailers — people who use platforms like Amazon to sell their own stuff. Once there, they’re unboxed, checked for damage, repackaged to Amazon’s standards, and sent on.

The Verge’s investigations editor Josh Dzieza discovered a cottage industry of these prepping centers that popped up in the small town of Roundup, Montana, an hour away from Billings. He wrote a story about how the tiny town became a hub in Amazon’s supply chain and talked about it with host Arielle Duhaime-Ross on this episode of the Reset podcast.

According to Dzieza, here’s the journey an item purchased online might take on its way to your doorstep:

“You can have a [third-party] seller anywhere in the world buy 100 dog beds from Target online. Those dog beds ship from a Target warehouse in San Bernardino, and then go to the prep center in Roundup, where they’re unboxed and re-boxed. Maybe Amazon says send 20 to Illinois, 20 to Kentucky, the rest of them to Utah. And then from there, the rest of them go out to customers.”

This system is so popular that there’s a very good chance your order will pass through Roundup — population 1,863 — this holiday season.

But why is it necessary for any item to go through so many layers of packing, unpacking, and repacking? “Amazon has fairly strict requirements [and a streamlined process] for how goods arrive at their fulfillment centers,” Dzieza explains. This extra step helps third-party sellers meet Amazon’s shipping demands.

Later in the episode, Jason Del Rey, senior commerce correspondent at Recode, sheds light on the history of third-party sellers, especially as they relate to Amazon’s business.

“In the year 2000, it actually started with big retailers who were the third parties. Toys R Us was one of the biggest back in the day,” Del Rey recalls. Over time, retailers built their own online stores, leaving small merchants and “one-person shops” to account for nearly 60 percent of total sales on Amazon today.

This system is lucrative for the online powerhouse, which charges sellers a listing fee, takes a cut of 8 to 15 percent, charges to store goods in Amazon warehouses (which is required to qualify for Prime), charges to handle customer service shipping, and charges to advertise these products on its site.

In a recent three-month quarter, Amazon generated $11 billion of revenue from it third-party services. 

“Without third-party sellers, Amazon would be much more like a traditional retailer having to stock all the inventory themselves. But this [current system] allows Amazon to have what might be the biggest online selection in the world,” Del Rey points out.

For more on how third-party sellers work on Amazon and to hear a conversation between Dzieza and the women who run Roundup’s prepping centers, listen to the entire conversation here.

Below, we’ve also shared a lightly edited transcript of Dzieza’s conversation with Duhaime-Ross.

Listen and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Josh Dzieza

So this started when I was talking to an Amazon seller for another story and he mentioned that he never really handled his goods. He bought them online from other retailers and had them shipped to a building somewhere where they were unboxed, re-boxed and sent to Amazon unboxing and re-boxed.

Arielle Duhaime-Ross

Why did they have to be re-boxed?

Josh Dzieza

That’s what I was surprised about, too.

Amazon has fairly strict requirements for how goods arrive at their fulfillment centers. Work in the fulfillment centers is partly automated and partly just intense physical labor. And to streamline the process, there are requirements for how things arrive. You can’t have multiple barcodes because someone might scan the wrong barcode, can’t have packing peanuts because they get everywhere.

So if you buy something from some other retailer, you have to unbox it, make sure it’s not broken, re-box it according to Amazon’s specifications, and send it to Amazon.

Arielle Duhaime-Ross

This process of unboxing and re-boxing to Amazon’s specifications — it’s called “prepping.” And Roundup is an unlikely hub for it.

Josh Dzieza

At last count there were nine prep centers, I think. And more people are training.

Arielle Duhaime-Ross

So all this unboxing and re-boxing activity in Montana, this cottage industry of preppers, how did this even come about? Josh went to Montana in September to find out.

Josh Dzieza

Prepping got started in Roundup because of this woman named Kristal Graham. Kristal’s brother died and left a bunch of books behind. 

To sell them off, she started selling on Amazon and then realized that you can sell pretty much anything on Amazon. So she started buying other things to sell — razors, KY jelly, first aid kits.

It soon came about that she had so many products that she couldn’t prep them all to send to Amazon. So she started looking around for services that would take that on.

Arielle Duhaime-Ross

That would prepare the packages for her? 

Josh Dzieza

Exactly. She went on the directory for seller services and saw that there were 15 such services at the time, but none in Montana. Sensing an opportunity, she decided to set up that kind of service herself. 

Kristal hired someone named Linda McAfee to help. We met at the Busy Bee Cafe, which is a diner when get into Roundup. 

Kristal had hired Linda McAfee and then they had a falling out. And Linda left and went into business for herself. And that was sort of where the prepping explosion in Roundup happened. 

She started getting things shipped to her property and packaging them and repackaging them herself. 

Arielle Duhaime-Ross

Okay. So now we’re at two prep centers in Roundup, Montana. 

Josh Dzieza

Then Jill Johnson hears about this from Linda’s neighbor. Jill had just been laid off from her state job and started trying to get Linda to show her the ropes. 

Jill started interning with Linda. Essentially, they became friends. And then Jill chose a name and started prepping from her property. So now there are three. 

And then Jill, who is the one who moved to Roundup to ranch, it came time for her to bring cattle down out of the mountains. She reached out to Sandi Green.

Sandi starts getting all these packages for Jill and she’s curious what the deal is. 

Arielle Duhaime-Ross

Because she doesn’t know what’s going on?

Josh Dzieza

She understood it to be some sort of mysterious online business but is curious what exactly is going on as the porch fills up with boxes.

Jill comes back out of the mountains and explains what’s happening. And Sandi thinks this sounds like a great flexible job.

Arielle Duhaime-Ross

So this is spreading through Roundup. 

Josh Dzieza

It’s spreading through word of mouth. And they’re just training each other. 

JILL JOHNSON: We have a group text. If anybody has questions. 

LINDA MCAFEE: They can go in the group, text a picture or ask a question. And you see me or Jill jump in there and answer them out. 

They were somewhat confused about why I was interested in this industry. Their sense is that this is a small, quirky thing they do. But they were quite warm and nice and eager to talk about the business.

A lot of Amazon sellers come and go, but some of their customers have been around for a long time and they have personal connections with these people who are all around the world and most of whom they’ve never met. 

LINDA MCAFEE: The majority of my sellers are international. I have them up in Malaysia, Australia, Europe, Canada. And a lot of them are starting out as newbies. Sometimes they need a little kickstart. They’re kind of like children when they first start, you have to play along. 

SANDI GREEN: It’s a lot of hand-holding.

Arielle Duhaime-Ross

I can’t help but notice that every person that you’ve mentioned in Roundup that does this prepping work, they all seem to be women. Why is that? 

Josh Dzieza

That is true. And that was something I asked them about. And they weren’t really sure why that was the case. They said it’s possibly because a lot of them were working from home, taking care of family. Some of them had husbands who worked in mines nearby. 

JILL: But you know what? We’re women that have our own businesses. What we’ve gotten all these women that are usually single. I mean, I have a boyfriend. She’s got a husband but he’s working in Billings and stuff. And, not many women get this opportunity. 

Josh Dzieza

They weren’t really sure. But it was particularly striking to me because talking to Amazon sellers, it can be sort of male dominated. It was interesting to me that this subculture, at least in Roundup, did tend to be women. It was entirely women in Roundup. 

Arielle Duhaime-Ross

And are people making money from this? This almost sounds like a multi-level marketing scheme. 

Josh Dzieza

They are. It is not super lucrative. They make $1 per package that they prep. It can fluctuate.

Sandi, for example, had a good day when she had hundreds of miniature animal toys to prep. At $1 each, you can prep them quite quickly. 

Arielle Duhaime-Ross

I feel like if it’s $1 per package, then you really want very small things. 

Josh Dzieza

Exactly. If you get stuck with a bunch of televisions or strollers or something, your hourly rate really plummets. 

They make good money. They said it was better than they can get at other lines of work.

Sandi did give me a rate for her miniature animal spree, which was $49.50 per hour for those. So quite good. But another day it could be a bunch of TVs, we don’t know. 

Arielle Duhaime-Ross

It’s not the most efficient system, is it? 

Josh Dzieza

No. One of the things about Amazon is that it is extremely efficient about the things that it focuses on optimizing. But because it’s so big, it can create weird inefficiencies like this of people trying really hard to capitalize on scale. 

Arielle Duhaime-Ross

It’s interesting to hear you talk about how Amazon just sucks in all these goods from various stores and companies across the country. It kind of feels like it has this special power. 

Josh Dzieza

Yeah. I mean, Amazon is this gigantic platform now that has made buying things extremely convenient and easy. And when you have a store like that, it can create this gravitational pull on consumer goods from all over the country of things flowing into Amazon fulfillment centers to be sold. 

And before they got there, they passed through a place like a prep center.

Listen and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

]]> Delia Paunescu <![CDATA[Why did Iran shut off the internet for the entire country?]]> https://www.vox.com/recode/2019/11/21/20975920/iran-internet-protests-reset-podcast 2019-11-22T09:10:12-05:00 2019-11-21T19:40:00-05:00
Iran imposed a nationwide internet outage after citizens flooded the streets to protest the government’s hike in fuel prices. At least five people were killed in the demonstrations, which show no sign of subsiding. | ATTA KENARE/AFP via Getty Images

Iran’s government announced on November 15, 2019, that it had plans to increase the price of fuel by at least 50 percent. The next day, Iranians took to the streets in protest. And the Iranian government reacted to its citizens organizing by shutting down the internet for the whole country.

The shutdown persisted for five days with practically every Iranian citizen lacking internet access or wireless data service. Finally, on the fifth day, Iran began restoring access in Tehran as well as in a number of provinces.

How is it even possible to shut down the internet for an entire country?

Lily Hay Newman, a senior security reporter at Wired magazine, explains on this episode of the Reset podcast that in order to get the internet to go down in the first place, the government had to get around a number of fail-safes.

“With this Iranian shutdown, which is so far beyond anything the government had done before, they likely grappled with these protective mechanisms. They would take down a portion of the network and then the network would automatically reroute around that dead portion to keep providing service. And then they would have to kind of do a whack-a-mole type situation to get it all under control.”

The “whack-a-mole situation,” as Newman describes to host Arielle Duhaime-Ross, “involves the government coordinating with internet service providers, telecoms, and infrastructure providers to actually make all of this happen.”

What this means for Iranians is that their government didn’t face a lot of resistance when it asked private companies to turn off internet access for their paying customers.

Alp Toker, the executive director of Netblocks, an NGO that monitors the governance of the internet, tells Duhaime-Ross quite frankly that he’s never seen such a “complex shutdown” before.

“We’ve been tracking a lot of shutdowns, and they are done for a lot of different reasons around the world. We always try to take the governments on good faith [and] see what they’ve done, whether there might be a justification. But, to be honest, in each case, it’s caused more harm than [good]. If it’s a protest or even a riot, switching off the Internet makes people more angry.”

Another alarming fact is that Iran might further isolate its citizens from the rest of the world through its own intranet, which the government has been developing in order to give the regime more control over what content and services people can access in the country.

According to Toker, a sovereign internet — like the one Russia was planning on building — leads to parallel internets and the loss of global connectivity.

But Newman argues that there might be an unintended benefit to the oppressive government’s total blackout. “Perhaps it could fuel even more people in the streets, going outside and saying, ‘The only way I can get information is to go talk to my neighbors [and] going to a protest.’”

Listen to the entire conversation here. Below, we’ve also shared a lightly edited transcript of Toker’s conversation with Duhaime-Ross.

This episode of the Reset podcast was published on November 21, 2019. Events may have changed.

Listen and subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Arielle Duhaime-Ross

So the internet’s out in Iran. And that means reporting on what’s going on within the country is a lot harder. 

It’s actually in these moments that I personally realize how screwed journalists around the world would be without the web. 

But there are some people out there who devote themselves to keeping track of these kinds of shutdowns — and they are a little bit more accustomed to navigating these communication obstacles and interpreting them. 

One of those people is Alp Toker, the executive director of Netblocks, an NGO that monitors the governance of the internet. He’s based in London, and I asked him if he’s seen an internet shutdown like this before.  

Alp Toker

Well, in short, we haven’t. 

It’s been extraordinary. It’s been a complex shutdown. It’s had many facets to it. It’s also impacted a lot of people at a time of crisis in various ways. It’s just unique. 

Arielle Duhaime-Ross

This shutdown was actually sort of gradual. The government initially just slowed down the internet for tons of people and then actually spent 24 hours just systematically trying to shut down the internet. What does that tell us about how the government is viewing these protests? Because this was a response to the protests. 

Alp Toker

Well, it stands to reason that the authorities are worried, given how extreme this measure is. It’s not something that Iran does. This is about severing connectivity at the core. 

Arielle Duhaime-Ross

Is there a link between the intranet project Iran is reportedly working on and the protests?

Please note: This is a developing situation, so I want to be clear here that while other news organizations have reported on what Alp is about to say, it’s still really hard to know exactly what’s going on in Iran.  

Alp Toker

So researchers knew that some kind of national intranet was coming. But there was no clear timeline. Would it be next year or a decade from now? And what wasn’t expected was that we might start seeing signs of the intranet during an internet shutdown that’s targeting protests.

Arielle Duhaime-Ross

Are we actually seeing signs of it being booted up now?

Alp Toker

All indications are that this system has been booted up. And that’s what we see from the data, from the reports, from all sources. What’s happened is: first this shutdown, but then you’ve got this network coming back online. But without the rest of the world there. So you’ve got a parallel intranet being formed in front of our eyes, where the people want to reach out but they can’t. 

Arielle Duhaime-Ross

This feels like big news, that Iran’s intranet is essentially its mode of communication. What we think of as the internet is essentially changing completely. That’s what you’re telling me?

Alp Toker

Right. Because now you’ve got to ask, which is a real internet? What if you have one of these or two of these? Is it the one inside Iran? Is it the one outside? What if Russia also does the same thing? Russia’s had some plans to [create] a sovereign internet and then you start developing parallel internets and you lose this global connectivity, this achievement of humanity, which has brought good but has also brought some perhaps less desirable things to the world. But that’s something we’ve all taken for granted until now. It looks like that’s changing. 

Arielle Duhaime-Ross

Do you think the government planned things this way? Was this premeditated? 

Alp Toker

It’s difficult to believe that this could be orchestrated. These protests were organic. They were caused by a specific trigger as the fuel prices rose.

A very cynical look at it might say that there was some planning. But honestly, it looks like the internet was cut. And then there was the realization that this is not good for the economy, this is not good for the situation of the country.

And [separately] there was this intranet that was being developed and it looks like that might have been switched ahead of schedule and might have enabled at scale.

So it remains to be seen how and why this is happening now. But it’s happened. 

Arielle Duhaime-Ross

Has this happened before in a country? Is there a country that has this kind of an intranet in operation and countrywide?

Alp Toker

The only similar high-profile country that has a system like this is North Korea.

Not much is known about the network that North Korea operates, but it doesn’t have many users. It’s specifically for those privileged [few] who can access this kind of network. And it isn’t connected to the global internet. Because it’s so isolated, it’s difficult to say how similar it is to what’s happening now. 

But you can look at Russia, which has been making plans for it but hasn’t activated this kind of system on a national scale. 

Arielle Duhaime-Ross

Have other countries in other parts of the world been affected by a government-controlled internet shutdown and communications blackout like this before? 

Alp Toker

A recent case is Iraq that has been cut off. First they switch off social media, then they switch off the internet completely.

And the human rights organizations are still stumped as to how many people have lost their lives. It’s created an information black hole that’s going to take years and there may not be an answer for. So that’s just one example. 

Arielle Duhaime-Ross

Is there ever an instance where shutting down the internet is acceptable?

Back in April when Sri Lanka suffered some terrorist attacks, the government decided to shut down its internet to stop the spread of misinformation. And actually the country got a lot of praise for that.

So is it ever okay to do this kind of thing and shut down the internet?

Alp Toker

We’ve been tracking a lot of extended shutdowns, and they are done for a lot of different reasons around the world. We always try to take the governments on good faith to begin with. We try to see what they’ve done, whether there might be a justification. But to be honest, in each case, it’s caused more harm than any problem it’s aimed to solve. If it’s a protest or even a riot, switching off the internet makes people more angry.

Let’s look at the situation in Sri Lanka. There you have these devastating terror attacks. Massive loss of life. And in the aftermath, there are disputes, groups that are blaming each other. And authorities introduce a blackout. They say it will prevent these attacks. They say it will stop these attacks. The problem is there is no evidence at all that it actually stops these attacks, because at the end of the day, people still know where to go if they want to attack a group of the minority. The problem is, it just stops the reporting. So nobody knows that it’s happening. And the problem is when you need it, when that press is really there when people are dying, then you’ve already lost the only recourse you have. 

Before we set up Netblocks, there were efforts to track internet shutdowns by hand or using various data sources. But a lot of this was missing — the information about how long it has lasted, what kind of disruption it’s been.

The truth is, we don’t know. We read stories, we read claims in archives that there was an internet cut. But that could have affected one city. It could have affected one street. And because the data wasn’t there. And that is lost to history.

Because you think if a country is shut down, the whole country, the world would know. But by its nature, there’s no way to record it. You don’t know if your neighbor’s offline. You don’t know if the next city is offline. You’re just out of information. 

Arielle Duhaime-Ross

This is actually the situation of if there’s a tree that falls in the forest and no one’s around to hear it, does it make a sound?

Alp Toker

Absolutely. This is the technical implementation of the falling tree in the woods. 

Listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[Google just got access to millions of medical records. Here are the pros and cons.]]> https://www.vox.com/recode/2019/11/19/20971337/google-medical-records-ascension-reset-podcast 2019-11-19T14:53:27-05:00 2019-11-19T14:43:26-05:00
The program, code-named “Nightingale,” is already creating major privacy concerns. | Photo by: Jeffrey Greenberg/Education Images/Universal Images Group via Getty Images

Google has been venturing into new areas of business and recently made huge news when it got access to the health records of millions of Americans through a partnership with the Ascension hospital network.

Both companies insist their goal is “to provide better care to patients,” but the program, code-named Nightingale, is already creating major privacy concerns. Just 48 hours after it was announced, federal regulators from the Department of Health and Human Services announced an investigation into whether the partnership violates HIPPA, the Health Insurance Portability and Accountability Act.

On this episode of the Reset podcast, Christina Farr, a tech and health reporter for CNBC, tells host Arielle Duhaime-Ross that this type of initiative isn’t exactly new for Google.

“Google has this history of walking into a new sector and and saying, “Let’s suck up as much data as we can, and we’ll just use our engineering prowess to figure out what products and tools that we can build off the back of this data.” And it seems like they’re going into health care with a similar intention.”

So, what does this mean for the Americans whose health records were accessed by Google?

The first thing to know is that when it comes to medical records, it’s not always clear who owns the data, says Nicholas Tatonetti, assistant professor of biomedical informatics at Columbia University.

“It is often generally owned by the doctors, hospitals, and organizations that collect it. But it seems to be a patchwork of regulations and frameworks, which is why it’s valuable for these large efforts to bring data together.”

For many Americans, that can be worrisome. But it’s important to know that the Google deal is actually routine. It’s part of a how a lot of medical research is conducted today.

Using an example from his own research — in which his team used similar types of databases to discover that two common drugs, an antibiotic and a heartburn medication, can lead to potentially dangerous heart arrhythmias when taken together — Tatonetti says that everything Google’s done so far “was perfectly above board.” But there is still room for improvement.

“There is a feeling in the air about who has access to our data, how private [is it], and when is it being shared? When we are left out of that process, it feels a little like we’re being taken advantage of, even if it’s legal. And even if they are appropriately protecting our data, I’m disappointed in these types of announcements [because] there isn’t an engagement of the patient population in order to bring them into this process, especially when it comes to a giant tech company that has the type of position and an ability to interact and reach billions of people.”

Still wondering how we should think about what Google is doing with these data and whether there’s a way to avoid all of the mistrust? Listen to the entire conversation here.

Below, we’ve also shared a lightly edited transcript of Farr’s conversation with Duhaime-Ross.

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Christina Farr

This blew up because of various reports that some of this information was not anonymous — that, in fact, it contained things about patients that you wouldn’t necessarily want to share with a company like Google. 

Arielle Duhaime-Ross

What kind of data was shared? 

Christina Farr

So the data that was shared was a bit of a mix. There were some cases where it was fully anonymized information, and that was simply to inform some of the analytics work that they were doing. In other cases, the companies came clean that they were sharing some personal health information which could have been potentially identifying. It could be all sorts of different things, including even just dates of service (when a patient went into the hospital).

We haven’t seen yet any clear evidence that patient names were shared in this process. But there’s a reason to worry about Google having access to any identifying information, because if they combine that with any other data they have about us, anything could be identifiable. 

Arielle Duhaime-Ross

We aren’t just talking about Google knowing people’s blood pressure, right? We’re talking about Google knowing people’s HIV status or whether they have a mental health issue that requires medication.

Christina Farr

Absolutely. The big fear here is that Google will start to learn more about our health conditions and [probably] already has quite a lot of that information.

We’re sharing our health status with Google inadvertently all the time. And the idea that they could then touch our clinical records from when we go see our physicians at the hospital is just terrifying.

Adding to this is that Google has seen other issues with some of the health things that it’s done.

Only a few months ago, there was a lawsuit from a patient at the University of Chicago. And what came out is that Google was supposed to be making sure that information that was shared from the university to their servers was fully anonymous. But it turns out that some dates of service … were actually shared with the company. So that led to a lawsuit.

Before that, there was a whole issue in the UK with DeepMind, one of their subsidiaries, having access to patient data.

This is all adding up to this picture that Google is not properly just managing this. I would like to see Google get out there and deny that and say, “We would never target people based on their medical information,” and just create some policies around this, and maybe even some public forums where people can ask questions of Google and get straight answers about how their health information is going to be used by the company. 

Arielle Duhaime-Ross

What about Google’s partner here? Why would they want to partner with a company like Google? 

Christina Farr

Ascension is a Catholic health system. They have lots of different hospitals and their own C suite that is looking for partnerships with tech companies, as are many other health systems. In the US, it’s quite common now to work with either one of the big three, whether it’s Microsoft, Amazon, or Google.

So, Ascension in theory would have wanted to have their brand associated with an innovative tech company like Google [be] very positive for them. I think they did see an opportunity to work with a big tech company on that and be viewed broadly as an innovative mover within the health care space. 

Arielle Duhaime-Ross

What was Google trying to do with the data? 

Christina Farr

My sources have said that there were a few projects that were outlined with Ascension specifically. One of them was that they were looking to build a tool that could search through a medical record really easily. I also heard that they were looking at early detection of disease. So, for instance, if a patient is likely to have a condition called sepsis, which could lead to a fatal outcome, is there a way that they could look at these large-scale datasets and figure out who’s most at risk and intervene earlier? And then once they had done that for something like sepsis, they could move on to other conditions. 

Arielle Duhaime-Ross

Is it normal for a free company like Google to have access to this data, especially if it’s not anonymous?

Christina Farr

These sorts of agreements are very common in the health care industry. Some folks in the wake of this news say that if these agreements didn’t happen, health care would grind to a screeching halt. So we see these deals all the time. Typically it doesn’t involve companies like Google, but it involves health care technology companies that you may be familiar with.

Optum, [for example], works with health systems regularly on large-scale data projects. And in these cases, they have to sign what’s called a BAA (or business associate agreement), which allows for this data to be shared and can actually include some personally identifiable data.

Google is just latching on to a long history of these preexisting types of projects that we see every day but [that] rarely get reported on. When it’s Google, it’s a big deal. But when it’s not Google, people don’t care quite as much. 

Arielle Duhaime-Ross

It sounds like, because the name Google is tied to this, people are reacting really strongly.

Christina Farr

That’s absolutely the case. This is standard practice. All sorts of health-privacy folks I’ve spoken to have said these deals are routine now. I think there is still reason to criticize Google.

One piece where I would call them out is consent. There was no evidence here that Google did tell any of the patients or the physicians that they were doing this work. That only came out later, after the news exposing some of the details of the project. They could have chosen to do that. Did they have to? Maybe not. Some BAAs allow for this to be done without consent.

I hope that once the dust settles, we end up having much deeper discussions about what we expect when it comes to our health information, who should own it, who should access it, and in what circumstances should patients have the right to say no. 

Arielle Duhaime-Ross

You alluded to a small number of Google employees who had access to this data. 

Christina Farr

Google hasn’t disclosed yet exactly who those employees were, who had access to the data. They said that this small number of employees were closely monitored, implying that those employees were watched if they did have any access to the information.

At this point, we just have to trust them that those employees weren’t sharing this data or attempting to use it for any nefarious purpose, like selling the data, trying to use it for targeted advertising, or even just building tools off the back of Ascension datasets that they could try to sell down the line to other hospital systems.

We don’t know that they did that. Google is asking a lot by just saying, “Hey guys, a few people had access to this, but don’t worry, we had it under control.”

Arielle Duhaime-Ross

It’s always fun when a large tech company like Google asks you to just trust them. 

Christina Farr

Exactly. As has been rightly pointed out, Google has been quite cavalier.

[The company] has this history of walking into a new sector and then saying, “Let’s suck up as much data as we can, and we’ll just use our engineering prowess to figure out what products and tools that we can build off the back of this data.” And it seems like they’re going into health care with a similar intention.

On the one hand, it could be a good thing, because this work does need to be done in health care and these large-scale analytics projects can be really important.

But on the other hand, you know, Google just hasn’t instilled the public with a lot of confidence that they’re going to approach this with all the protections and the controls and just fundamentally do it in the right way. 

To hear the potential positive aspects of Google’s partnership with Ascension and its access to massive amounts of medical records, listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[America’s sex education system is broken. This chatbot wants to be the solution.]]> https://www.vox.com/recode/2019/11/17/20965326/sex-education-planned-parenthood-chatbot-roo-reset-podcast 2019-11-18T16:19:19-05:00 2019-11-17T09:32:39-05:00
Planned Parenthood created Roo, a sex ed chatbot that answers teens’ questions while maintaining their anonymity. | Planned Parenthood

US sex education is decades behind other countries. Right now, it isn’t even mandatory in every state. Add to that the awkwardness people feel about sex and bodies in general, top it with the idea of having to have these discussions in public, and what you get is a system that’s devolved into a total mess.

It’s no wonder 84 percent of teens look for health information online. The problem there is that a lot of the answers they come across — about everything from STIs to puberty to pregnancy to sexual orientation — are often just plain wrong.  

Naturally, tech wants to find a solution. 

That’s how Planned Parenthood created Roo, a sex ed chatbot that encourages teenagers to ask all of their potentially uncomfortable sex-related without ever revealing their identity.

To build Roo, Ambreen Molitor, senior director of the Digital Product Lab at Planned Parenthood, first interviewed Brooklyn high school students about their online habits and what they would want out of a bot that talked to them about everything from safe sex to coming out. Her team discovered that above all, “teens really wanted to be anonymous.”

“Sometimes they didn’t feel comfortable talking to the community around them or in the sex ed classrooms. But also online, because more often than not, Gen Z’s teens in general are very aware that when you’re searching on Google, you’re being cookie’d. They’re very cognizant of what they type into the browser or the search query — which is really unique.”

In this episode, Molitor tells host Arielle Duhaime-Ross that Roo is seeing great success so far. Parents have even reached out on LinkedIn to praise her bot.

Of course, the complexities of human sexuality, specifically as they need to be explained to a developing and curious teenage population, can never be fully resolved through an anonymous computer that’s been preprogrammed with answers.

That’s where Nora Gelperin, a parent and longtime sex educator who’s currently the director of sexuality education and training at an organization called Advocates For Youth, comes in. She developed a sex ed video series called Amaze.

With over 80 installments on topics ranging from gender identity and sexual orientation to sex trafficking, intersectionality, puberty, and even wet dreams, Gelperin revealed that technology can be “really a great companion for adults, whether they’re parents and caregivers or professionals having these conversations.”

But don’t expect a chatbot like Roo or even an extensive and informative video series like Amaze to solve the problems that a lack of comprehensive sex education leave behind.

“I think that there is a lot of information that needs to be supplemented to any of those technology-based resources because they can’t talk about values, they can’t talk about what do you do if you think you want to have an abortion but your religion tells you you’re going to go to hell. Or what do you do if you think you’re committing a sin by masturbating. Those are the things where the technology is kind of limited,” Gelperin says.

Listen to the entire conversation here, where you can find out what a high-school-age person actually wants to be told about sex. Below, we’ve also shared a lightly edited transcript of Molitor’s conversation with Duhaime-Ross.

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Arielle Duhaime-Ross

Kids and teens are hungry for this kind of information.

Ambreen Molitor

Almost 84 percent of teens actually look for sexual health information online. So our team built a sex ed chatbot named Roo. It’s only 9 months old. Very much in its infancy. 

Roo allows folks, specifically teens, to anonymously ask all kinds of questions around sexual health information. The interface is very much like a text format. So Roo will prompt you, greet you, and allow you to have the open space to ask a question. It can be as short or as long as you want and Roo will respond to you in 180 characters or less.

Arielle Duhaime-Ross

I’ve had pretty terrible experiences with chatbots and they don’t have the greatest track record in general. Maybe you remember Microsoft’s chatbot attempt a few years ago. They had to shut it down because Twitter managed to train it to be racist and misogynistic in less than 24 hours.

So when I heard about Roo, I was honestly pretty skeptical. I know Planned Parenthood is good at teaching people of all ages about sexual health but I wasn’t convinced the organization would have the tech chops to make a bot that didn’t suck. So I decided to put it to the test.

My experience with Roo wasn’t terrible. And that’s surprising. So I asked Planned Parenthood how they went about designing it. 

Ambreen Molitor

How Roo works is three-fold. 

First there is software that is built. It’s artificial intelligence, and the actual software that we use is called Natural Language Processing (NLP). For folks who are not familiar with what that does, it’s the same software that allows you to talk while you’re texting, it completes your word or completes your sentence.

That’s the same software we’re powering with Roo. So Roo is trained to anticipate the question and also anticipate the sentiment of the question to be able to answer it. 

The second and third layer are human inputs.

The second input is we have a content strategist that comes in and ensures that the answers that we provide have that nonjudgmental tone. It provides the personality that brings Roo to life.

The third most important one is a team of educators that reviews each answer and ensures that it’s medically accurate and up to date.

Arielle Duhaime-Ross

Ambreen’s team talked to teens at a high school in Brooklyn about their online habits and what they wanted out of the bot.

Ambreen Molitor

Teens really wanted to be anonymous. Sometimes they didn’t feel comfortable talking to the community around them or in the sex ed classrooms, but also online. More often than not, Gen Z’s teens in general are very aware that when you’re searching on Google, you’re being cookie’d. So they’re very cognizant of what they type in the browser or search query, which is really unique.

Arielle Duhaime-Ross

Planned Parenthood actually even has a texting service where you can talk to a sex ed professional directly. But now you’re developing a chat bot. So it sounds like you still feel a need to remove a human from the equation even further.

Ambreen Molitor

Yeah. That’s because at certain times we found that teens feel comfortable with talking to a bot because it eliminates some strong bias and they’re quick to open up to the actual questions they need to get to.

Arielle Duhaime-Ross

According to Planned Parenthood, teens like using Roo because it protects their anonymity. And the fact that it comes in the form of a cute little avatar doesn’t hurt.

Ambreen Molitor

It’s gender-neutral. You cannot determine if it has a certain gender identity or even sexual orientation. If you take a long time to type something, Roo starts to like fall asleep and has some Z’s going over his head. And they love that. They’re like, “This avatar is actually paying attention to me. They’re taking the time to understand and connect with me in unique ways.”

Another thing we get so much feedback on is, “Not only is it great that I feel safe, but I also feel like this avatar really is listening and understanding my habits.”

Arielle Duhaime-Ross

Does it have a human form or human shape?

Ambreen Molitor

It’s a blob. It’s just basically an avatar that’s a rounded rectangle with eyes and a mouth to provide gestures.

Arielle Duhaime-Ross

So it’s your friendly neighborhood sex ed blob.

Ambreen Molitor

Precisely. It winks, sleeps, snores, all of those things.

Arielle Duhaime-Ross

Planned Parenthood doesn’t keep track of who uses Roo but users can opt in to share information about their age and race.

Ambreen Molitor

Of those people who opt in to provide that information, 80 percent of them have identified as teenagers. So it’s about 60-40 percent male to female and 2 percent other gender identities.

Almost 70 percent of the folks that we talked to — again, who have opted in to provide us information — are what we consider people of color. So they’re of a diverse background and race and ethnicity.

Arielle Duhaime-Ross

Planned Parenthood also monitors the questions people ask Roo. And some of the subjects teens broach with Roo have been surprising. 

Ambreen Molitor

Consent is a topic that we did not anticipate either from the learnings through visiting the high school or through the data that we were seeing from our website. Otherwise, we anticipated lots of questions around puberty and around those changes.

Arielle Duhaime-Ross

“Is this normal …”

Ambreen Molitor

Correct. The spectrum of normalcy is what every teen wants to understand, it’s where they live. Normal is very important to a teen. And that’s something we knew going into it.

Arielle Duhaime-Ross

These are big, complex, heavy topics. How does Roo answer these questions in 180 characters?

Ambreen Molitor

We start off by describing consent. We say that there’s no one way to do it. And then we provide just an example or some guiding principles for that. Once we answer the question, we recognize that someone may want to go deeper. And we have link-outs to pages on our site and videos on how you can find or figure out different ways that people can ask for consent. So it goes one step deeper when 180 characters cannot fulfill the curiosity that someone has about that question.

Approximately 80 percent of the time, we’re answering the question correctly. A lot of it falls on two years worth of data and testing that we did. So we didn’t just launch it and go with it.

The other reality we need to call out is that machine learning is not 100 percent accurate. I think Roo’s very humble to say, “I’m not built to answer this question,” or, “I don’t understand it,” or, “I actually don’t think it’s appropriate for me to answer it.” And we’re really good about handing it off to a human. 

Arielle Duhaime-Ross

So there are questions Roo can’t answer. Which means Planned Parenthood had to build in some guard rails. For instance, if someone appears to be in crisis, Roo will hand off the conversation to a mental health hotline.

Ambreen Molitor

The other time that Roo does handoffs is when there’s decision-making in mind. So the birth control question is a really good one where there are several different birth control methods and there’s not one directional way to suggest this birth control method that’s universally great. That’s where decision-making comes in. That’s an opportunity for Roo to understand that it’s best to hand it off to an educator.

Arielle Duhaime-Ross

The feedback from teens seems to be positive so far. But there’s another demographic that Roo has also been attracting.

Ambreen Molitor

It’s so funny. Parents love this. I’ve actually had, anecdotally, parents reach out to me on LinkedIn and say, “Thank you so much for this bot.”

Arielle Duhaime-Ross

On LinkedIn — great place to talk about sex ed.

So Roo seems to be a surprisingly not-terrible chatbot. But when I think about Roo, I honestly feel kind of sad. Because I see why teens might prefer to use Roo rather than turn to an adult for help. And that makes me wonder:

Why is sex ed so broken? And is Roo really the solution?  

For the answers to these questions and many more, listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[A bait-and-switch scam ran unchecked on Airbnb. Here’s how it worked.]]> https://www.vox.com/2019/11/14/20961972/airbnb-scam-how-to-stay-safe-reset-podcast 2019-11-14T15:07:35-05:00 2019-11-14T12:30:00-05:00
While Airbnb has been slow to implement strong verification policies, some customers are getting scammed in a big way.  | Photo Illustration by Avishek Das/SOPA Images/LightRocket via Getty Images

While on vacation in September 2019, reporter Allie Conti accidentally uncovered an elaborate Airbnb con when she herself became the unwitting victim of the scam.

Conti and her friends paid Airbnb $1,200 to stay at an apartment in Chicago. But just as they were about to check in, the host called to say the place was flooded. He offered another property of his instead but that alternative turned out to be pretty dingy. Even worse, the host kicked out Conti and her friends after just two days — forcing them to find a hotel at the last minute.

The host never offered them a refund. But Airbnb ended up refunding her in full after her article was published.

When Conti researched her horrible host, she found that the same people responsible for scamming her also managed Airbnb listings in eight cities across the US.

On this episode of the Reset podcast, she tells host Arielle Duhaime-Ross how the scam works:

“You book a place that looks pretty nice and pretty cheap and then about five minutes before check-in you get a call about an emergency. It’s flooded, the air conditioner’s broken, something insane is happening. You don’t want to stay here. But luckily, he’s there to save the day with a nicer, bigger place. You’re going to be disoriented in a new city and just decide to do the easiest thing, which is to trust this person that you don’t know. You get it there and it’s just disgusting. But you have to request a refund before check in. So it’s too late at that point to ask for a refund through their policies as written.”

Later in the episode, New York Times tech reporter Mike Isaac breaks down what about big tech companies makes them ripe for exploitation at the hands of crafty criminals. Isaac explains that Silicon Valley’s insistence on “scaling up” and having companies like Facebook, Uber, and even Airbnb grow very big, very fast leaves little room for them to focus on “doing the proper due diligence to make everyone safe and happy.”

“That’s just the nature of being a platform. You have to get big before you can police your content. And then it becomes impossible to properly police it once you get that big. So it’s kind of a Catch-22.”

If you’ve ever rented a place or Airbnb or plan to one day, listen to the entire discussion here. Below, we’ve also shared a lightly edited transcript of Isaac’s conversation with Duhaime-Ross.

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Arielle Duhaime-Ross

The scam that Allie discovered is actually just one of the scandals Airbnb is dealing with right now. How could these issues have been prevented? And why weren’t they? 

Mike Isaac, you just came out with a book about Uber, which is perfect because Uber, like Airbnb, is an app that connects customers to people who own a thing like a car or an apartment that they can share.

So, Mike, walk us through the last couple of weeks for Airbnb?

Mike Isaac

It’s just been a series of total press PR nightmares for Airbnb the past few weeks. They had this really awful tragedy happen in this community called Orinda in Northern California, where folks had rented out an Airbnb house to basically use it as a party house for Halloween.

Some really awful shootings happened and people got killed over what should have been like a fun night. Folks essentially said that Airbnb should be more responsible in policing that.

And then Allie Conti stumbled upon this crazy widespread scam where folks are essentially getting ripped off for renting Airbnbs and then getting duped last minute when they got to the place. It was a way of scamming people out of paying a lot more than they should for these properties.

So all of that kind of blowing up at one time for the company has not been great on showing how Airbnb handles these properties. 

Arielle Duhaime-Ross

How has the company responded? 

Mike Isaac

They’re doing these standard corporate things — “We’re looking into this. We’re taking this seriously. We guarantee we’re going to police our platform a lot better.” They said they’re going to start verifying each property and making sure what is advertised is the correct thing that you’re going to get. But I see it as kind of a little bit of lip service. 

Arielle Duhaime-Ross

So Brian Chesky, Airbnb’s CEO, recently said in an interview with Kara Swisher that his company had been slow to implement strong verification policies.

Mike Isaac

That’s kind of a head fake, honestly. Part of the whole philosophy of becoming a platform — and Airbnb is one of the largest platforms for home sharing and renting your place in the world — is expanding as fast as possible. That means just getting people to sign up and list their houses or apartments or whatever. By design, those verifications and checks in the process are not going to be built in from the beginning just because you have to have what’s called liquidity on the platform. You have to give people as much selection as possible.

The way that technologists view it is some subsection of our properties are always going to be false or at least not properly vetted. That’s the sort of percentage that our platform is willing to deal with in order to make this work in the long run.

So I would argue that it’s kind of built by design that way initially. And then later on, once you get to a big enough scale, then they could say, “Oh, we’re going to start doing the proper due diligence to make everyone safe and happy.”

Arielle Duhaime-Ross

These policies that would be designed to keep people safe, to make sure that there’s a very low percentage of scams on the platform, they cause friction. They stop people from signing up. They’re a barrier to entry. And companies like Airbnb don’t really like that. 

Mike Isaac

Yeah. Hundred percent. All of these are very interchangeable. Airbnb might have said we could have had better vetting policies upfront, but that’s the same as Uber, which has gone through this process of background checks for drivers and making it harder for people to sign up for the platform if they didn’t have identity verification.

That all comes later once they get to the scale that they are early on. Uber was just kind of taking as many bodies, whether as riders or drivers on the platform as possible, just as Airbnb was trying to do.

They’re like software corollaries to this, too. You can look at YouTube in its earliest days. All they wanted was just getting as much video content on the platform as possible to grow. And that included copyrighted material that was just ripped off or was free-booted material from other networks.

That’s just the nature of being a platform. You have to get big before you can police your content. And then I would also argue that at the same time, it kind of becomes impossible to properly police it once you get that big. So it’s kind of a Catch-22. 

Arielle Duhaime-Ross

What is it about that culture that these companies respond only after something bad happens and a reporter writes about it? Is this a Silicon Valley thing? 

Mike Isaac

This is why I take all of these “We’re very concerned” statements with a real grain of salt because everyone who’s building these platforms knows exactly what they’re doing. And sort of by design — this is how it was meant to scale.

It’s fair to argue that you don’t know exactly how the platforms are gonna be exploited because criminals or thieves are very creative. And we’re finding new ways that Facebook is being manipulated every day. It’s hard to really predict how your platform is going to be used.

That said, there’s an acceptable amount of risk built in to building any of these things in the first place. It’s really about getting to scale and doing that as quickly as possible for it before some other competitor beats you to it. 

Arielle Duhaime-Ross

I feel like these companies want to get as big as possible and as quickly as possible. Why, though, is it so important for these companies to reach scale? 

Mike Isaac

There is the advantage of what’s called “a network effect,” which means the bigger you get, the more your platform reinforces its own entrenched incumbency.

Arielle Duhaime-Ross

You’re the dominant force, basically.

Mike Isaac

Facebook greatly benefits from that. The more people that use it, the more people are going to continue using it over time. Part of that is maybe the nature of the business itself.

I can’t imagine most of the folks in Silicon Valley wanting to be fine with a small, modest, and sustainable business that isn’t growing by 100 or 200 percent every other quarter. It’s just about changing the world and world domination.

For the past 15 or 20 years, that was a lauded approach to how we look at CEOs. Now that sort of view is becoming questioned in tech, which is in for this reckoning right now. 

Arielle Duhaime-Ross

It’s interesting because I think that for a lot of people listening to my question, they would just go, well, money, obviously. But clearly, it’s not just money, it’s more than that?

Mike Isaac

I’m hesitant to put all this stuff on money just because a lot of the guys — and again, it’s mostly guys that are running these companies — have money right there. They’re set for life. Zuckerberg isn’t doing it for the money. He has more money than you will ever be able to spend.

It’s about conquest. It’s about history. It’s about making one’s mark on the world. Or if you want to go, Steve Jobs, the dent in the universe thing.

Really, it’s about ego.

I’m hesitant to say these businesses only care about money because I think it’s not the thing that really drives them at the end of the day. It’s about maintaining power and and beating your competitors so that you don’t have to worry about being usurped or becoming irrelevant. 

Arielle Duhaime-Ross

And in order to do that, you say I’m okay with 2 or 3 percent of listings being fraudulent and customers be damned. That’s just their problem. 

Mike Isaac

There’s always going to be laws being broken or crimes being committed or some “acceptable” amount of problems. It always exists.

So it might be hard to hold the platforms to a 100 percent standard of keeping everything pristine. But they also know what they’re getting into when they’re building as quickly as possible without proper verification.

So you have to come in and say, look, you need to forget worrying about growth at all costs and start protecting people on your platform, especially when it’s having real world, potentially life-threatening consequences.

To find out, listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[Why is SMS texting a mess? Fixing it is harder than you think.]]> https://www.vox.com/recode/2019/11/12/20960865/sms-texting-iphone-android-rms-reset-podcast 2019-11-12T12:13:40-05:00 2019-11-12T11:30:00-05:00
Ben Affleck was seen texting on an iPhone 11 Pro in Los Angeles recently. | BG004/Bauer-Griffin/GC Images

On November 7, 2019, more than 168,000 people woke up to find that text messages they thought they’d sent on Valentine’s Day in fact were only just delivered overnight, nine months later. The snafu became a nightmare scenario for many.

Blair Hickman, Vox’s director of audience, was one of those people. On this episode of the Reset podcast, Hickman tells host Arielle Duhaime-Ross how she found out her February text had only just arrived to her now-ex’s phone at 3 am the night before.

So, Thursday morning, I was out walking my dog, and I got a text from my ex — and we haven’t talked since June. All the text said was, “Ha ha. I don’t think this was for me, but I hope it was at least a pumpernickel bagel.” It’s like, I don’t understand. I hadn’t texted him, so it was just out of nowhere.

I said, “I don’t know what this is about.” He sent me a screenshot of his text messages, and there had been a text from me at 3:36 am that morning, and it said in all caps “AND A FUCKING BAGEL” with like six exclamation points.

The confusion shed light on the fact that the SMS (short message service) texting technology we use every day — and more than likely take for granted — is actually quite old as far as technologies go.

Carriers rely on third-party vendors to actually deliver the messages. And in the case of the Valentine’s Day glitch, a company called Syniverse, which provides networking services, took the blame for the messages being delayed. They say it happened because a server malfunctioned on February 14. Then when the server was fixed on November 17, all of the trapped messages were finally sent out.

It’s because of trouble like this that many consumers are switching to richer messaging services like WhatsApp, iMessage, Signal, and even Facebook Messenger. But the truth is, SMS is universal. Because it works on nearly every phone, regardless of data capabilities, it’s become the most popular messaging service in the world.

Dieter Bohn, executive editor at The Verge, explains exactly what the technical issues with SMS technology are both from a user perspective and from the viewpoint of the app itself:

“Text messages are such a primitive technology that’s just unencrypted text going out over the network. It runs through servers and different carriers have different policies about how long they hang onto those messages. In theory, all those messages should have just like gone into the ether, but in practice, they got stuck on a server. …

What everybody wants is to get on to some sort of universal system that actually is more secure than SMS but also has all of those [modern messaging] features.”

One possible fix for these issues is a technology called RCS (rich communication services), but it’s not going to be perfect. Bohn says that although RCS is a much better experience, it probably won’t be encrypted, creating privacy issues.

“[Carriers will] be allowed to potentially see the images that are sent. They could see what’s going on in group texts. To be clear, there’s no technical reason that RCS couldn’t be end-to-end encrypted. They’ve just chosen not to,” Bohn pointed out.

Later in the episode, Lloyd Cotler, co-founder of Banter Messaging — an agency that helps organizations talk to customers through SMS — reveals what it was like to be the first-ever SMS director for a political campaign when he worked for Democratic presidential nominee Hillary Clinton in 2016.

It’s basically a direct communication from either the candidate or representative of the campaign to voters who are supporting them or want to learn more. So folks kind of text in, they join the list the same way they would go to a website and fill out their email or sign a petition. You ask them their name, address, zip code. And then basically, you just start texting them.

Cotler praises text campaigning for allowing candidates to reach constituents they might not normally be able to find on the ground, whether it’s because they live in remote places, don’t answer phone calls, or are deaf or hearing-impaired.

SMS is really just the most common denominator that everybody can be a part of. Think about who you text in your own life — it’s your friends, your family, people that you’re very close with. So if an organization can achieve that level of intimacy with you, I think that’s great.

But how will the change to RCS texting impact political campaigns and their constituents in the future?

To find out more about the trouble of SMS technology and what to expect from texting in the future, listen to the entire discussion here. Below, we’ve shared a lightly edited transcript of Bohn’s conversation with Duhaime-Ross.

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Dieter Bohn

I was really surprised SMS has been around for so long. It’s the sort of thing that you think you can just take for granted. But it turns out that it is way more complicated and complex than we realized.

Arielle Duhaime-Ross

So why did this Valentine’s day SMS snafu happen?  

Dieter Bohn

When you send a text message, your phone carrier doesn’t necessarily know where it’s supposed to go just by from the phone numbers. They have this thing called an interconnect word, like all the carriers talk to each other. And there are companies that sit in the middle and handle that.

One of these companies is called Syniverse. They they claim to handle something in the neighborhood of 600 billion messages a year. And they are one of the companies that make sure that an SMS goes from one place to another.

This company had a server malfunction on Valentine’s Day, of all days. And instead of the messages just disappearing, [nine] months later they went to go check out the server, got it fixed, then all of sudden all these messages went out.

Arielle Duhaime-Ross

Most people don’t think about their SMS text messages being shared with anybody other than the recipient and maybe the carrier, right? 

Dieter Bohn

That’s what you expect. But in fact, text messages are such a primitive technology that that’s just unencrypted text going out over the network. So it just runs through servers and different carriers have different policies about how long they hang onto those messages. In theory, all those messages should have just gone into the ether but in practice, they got stuck on a server.

Arielle Duhaime-Ross

What does that tell you about privacy and SMS?

Dieter Bohn

Privacy in SMS is not great. But privacy in general with cell phone carriers, especially in the US, is also not great.

They are more than willing to respond to sorts of questions from the government, whether they’re subpoenaed and sometimes if not. They maintain lots of metadata, which is the information on who you talk to and when.

And some [carriers] do hang on to the actual content of your text messages on their servers for a while and can serve those up to the government if asked.

Arielle Duhaime-Ross

What did the carriers say when all of these stories started popping up about Valentine’s Day texts that are months old? 

Dieter Bohn

They really didn’t want to get into it [or] reveal the complicated ways that these companies actually have to interact with each other.  

Arielle Duhaime-Ross

Obviously, this Valentine’s Day thing has taken up quite a bit of room sort of on the internet. People are talking about it. But do you think this kind of thing happens more often than we know about?

Dieter Bohn

If it happened this regularly at this scale — the most recent estimate was it was 170,000 messages, which adds up to over 300,000 people affected by this — that’s a lot.

It probably happens on a much smaller scale more often. But we don’t have numbers for it. And the carriers, given their terse response to this, certainly aren’t interested in sharing it with us.

Arielle Duhaime-Ross

Why doesn’t SMS work very well all the time?

Dieter Bohn

SMS doesn’t work very well all the time because it often runs over the cellular network. So they just can’t build all the extra information in that you’d want from modern messaging service.

You don’t have a read receipt, so you don’t know if your message was received. You can’t send photos. Group chats turn into hellacious disasters all of the time. So everything that feels like it’s a modern messaging experience there is actually just sort of a hack that was built on top of SMS.

What everybody wants is to get on to some universal system that actually is more secure than SMS, but also has all of those other features.

Arielle Duhaime-Ross

Are there any plans to make SMS better?

Dieter Bohn

There is a plan to make SMS better. It’s been around a long time. It’s called RCS, which stands for Rich Communication Services. Google’s been trying really hard to have this adopted, because then [Google’s phone software] Android would have a good text messaging app by default. But a lot of the carriers have been dragging their feet on adopting it because it’s a lot of extra work to implement this thing. And SMS has been treating them just fine up until now.

Arielle Duhaime-Ross

That seems like a big deal. What does that actually mean? What is RCS? 

Dieter Bohn

It follows that same model as SMS where there’s different carriers that control their own servers but they do a better job of talking to each other. And it’s more advanced, which means that you can get most of the modern features you expect from a text messaging app like reader seats, like high quality photos, videos, and really good group chats.

It’s a much better experience so long as the phone you’re using and the phone you’re talking to both have it turned on by their carrier. Up until relatively recently, that’s been a crapshoot. It’s actually still a crapshoot. But starting [in 2020], the four major carriers in the US promise that they’re going to make it universal.

Arielle Duhaime-Ross

Who are the carriers that are involved with this?

Dieter Bohn

Sprint, T-Mobile, Verizon, and AT&T.

It’s called the CCMI, Cross Carrier Messaging Initiative. It’s a joint venture or two completely separate companies that all four of these companies have agreed to work with. It’s going to spin up the RCS server and it’s going to make sure that all of these phones that are on these networks are going to be old to have this more advanced messaging protocol.

Unfortunately, it also is probably going to mean that a lot of Android phones are going to have an app that’s made by the CCMI. And that’s troubling because historically the apps that carriers make are — what’s the word — terrible?

Arielle Duhaime-Ross

Given that track record, how hopeful are you that RCS will actually be something that people want to use and should use?

Dieter Bohn

I am more hopeful than you might expect. For one, Google — which controls 80 to 90 percent of the smartphones on the planet, that’s how big Android is — they have a vested interest in making sure that this is not a crappy experience because the crappier the default texting experiences on Android, the more likely people are going to go to get an iPhone. So they hopefully will put some pressure on the carriers to do this right.

The most crucial thing that people should pay attention to with the advent of RCS is whether or not anybody starts saying anything about encryption and specifically end to end encryption, because that’s the kind of encryption that guarantees that only the sending phone and receiving phone can decrypt the message and see what’s on it. 

Arielle Duhaime-Ross

Right now it looks like RCS probably won’t be encrypted. Is there actually a chance that carriers will be able to get even more information from users through RCS? 

Dieter Bohn

There’s definitely a chance they could get more information from users using RCS. They’ll be allowed to potentially see the images that are sent. They’ll know when read receipts went back. They could see what’s going on in group texts. It’s unlikely the carriers are actually going to be like actively reading your conversations. to be clear, there’s no technical reason that RCS couldn’t be end to end encrypted. They’ve just chosen not to. 

Arielle Duhaime-Ross

SMS will keep existing at least for a while. Is that important? 

Dieter Bohn

It matters insofar as as sort of the universal fallback. There are a bunch of things that can fall back to SMS if an IP-based data message doesn’t go through. Some emergency alerts, if it can’t get out over a data connection, might end up getting sent over a text message. And to be blunt, that’s one of the reasons why I’m actually pretty angry about this Valentine’s Day text message screw up.

If you’re not sending text messages — the thing that we all believe we can fall back on and rely on — you need to be much more transparent about that, because it’s one thing to get a text from your ex, it’s another thing to not get a text from the local fire department or your doctor or something else.

Arielle Duhaime-Ross

So depending on your phone carrier, the default texting app that you use could change because of this upgrade and it’ll let you send richer text messages — that could also further compromise your privacy. Funny how that happens, right? 

And in some ways the upgrade makes sense because as Dieter mentioned, the tech that powers text messages is old and definitely due for an update. 

But here’s the thing: Plain old dumpster-fire SMS text messages are actually super-important technology. 

To find out how SMS texting impacts political campaigns, listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[Wildfires are getting worse. Can new technology make us safer?]]> https://www.vox.com/recode/2019/11/10/20955996/california-wildfire-prevention-fire-technology-reset-podcast 2019-11-10T09:01:44-05:00 2019-11-10T09:10:00-05:00
Firefighters work near Getty Center in Los Angeles to control a fast-moving wildfire on October 28, 2019. Thousands of residents were forced to evacuate their homes. | Qian Weizhong/Xinhua News Agency/Getty Images

Wildfires are getting more intense. In fact, 2018 was the worst recorded year for wildfires in California with almost 2 million acres burning down.

Part of this is caused by climate change, which is happening all around us. You might say the escalating wildfires are a giant SOS signal asking us to pay attention.

But the good news is that new technology is being developed to help authorities find the blazes and keep people safe.

Graham Kent operates ALERTWildfire, a network of nearly 400 cameras around California that accurately tracks wildfires. He also has cameras installed in Oregon, Nevada, and Idaho, all of which are remote-operated and installed on mountaintops and in towers so they can be easily pointed at a source. “The whole idea behind it is to give somebody at dispatch the ability to see that fire right away so that they can scale up or scale down resources.”

Kent estimates that, in a recent fire, his cameras probably saved firefighters around 30 minutes, which can make a huge difference when it comes to stabilizing a rapidly spreading fire. 

Back in 1910, a large fire known as the Big Blowup razed 3.25 million acres of land in Idaho, Montana, Washington, and British Columbia in just two days. To battle the flames, 10,000 people went out using rakes, shovels, and two-man saw axes. 

These days, camera networks like ALERTWildfire along with drones, satellites, planes, and even handheld GPS scanners help detect wildfires so fire crews can get them under control significantly faster.

But getting firefighters to adopt new technology hasn’t always been easy. Sean Triplett with the US Forest Service explains that it has taken a lot of work to even get this far. 

“You have to be cautious with technology because if we try to deploy [new] technology in an operational perspective and it fails or has a negative outcome, you’ll have a very hard sell the next time around,” he says.

And that makes sense. “Firefighting is risky. And when you’re out in the field, you have to be able to trust the tools you’re working with,” Reset host Arielle Duhaime-Ross adds.

But as firefighters adopt new technology — however cautiously — who and what is being replaced? 

Later in this episode, singer/songwriter Marina Avros shares her experience as a fire lookout in Canada, a job she’s done every summer for the past 12 years.

“We’re part of this network of fire detection,” she says. “I feel like I’m a part of a bigger picture or a piece in the puzzle. So if they took us away, they would be losing a really important factor in early fire detection.”

Listen to the whole conversation here. Below, we’ve shared a lightly edited transcript of Avros’s conversation with Duhaime-Ross. 

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Arielle Duhaime-Ross

Humans still have a role to play [in fire detection]. And really, the job of a fire lookout is unlike any other.

Sean Triplett

You’re up on top of this mountain. Ten thousand [feet], sometimes even higher up in elevation and you just have these great panoramic views. But except for certain spots, fire towers and fire lookouts are on their way out in the US. 

Arielle Duhaime-Ross

Now, I actually know a fire lookout in Canada. So I called her up to find out what the job is like. 

Marina Avros is a musician who goes by the stage name of MarinaMarina. I met her a few years ago at a concert she was playing in New York City. We hadn’t spoken since — but when I started looking into fire detection tech, she came to mind. 

Turns out, she’s spent the last 12 summers in a lookout tower in Alberta, Canada. When I spoke to her, she was at her tower. 

Every day, Marina wakes up in a cabin on the ground. Then, she climbs 100 feet up in the air to watch the forest from a box about the size of a small bathroom.  

Marina Avros

Honestly, I am actually pretty scared of heights. There will be weeks where I can’t even look out my window, like to look down to the ground.

Arielle Duhaime-Ross

Most days, Marina just watches the forest. But a few times a season, she’ll spot something in the distance: a line of smoke. When that happens, she uses a tool called an Osborne Fire Finder. This is old tech. It was invented around the time of the Big Blowup, in the 1910s.

Marina Avros

It’s a metal disc. It’s maybe 2 feet in diameter. It has a map on it. And the map is of the area that I’m looking at. And then on top of that, there is a scope.

Arielle Duhaime-Ross

When Marina sees smoke, she lines her scope up and takes a measurement of the angle, basically drawing a line between her and the fire.

Other fire watchers nearby do the same and share their info. Where the lines cross, that’s the location of the fire. 

There are about 125 lookouts in Alberta. They’re basically a retro version of Graham Kent’s camera network. 

Marina is actually part of a long tradition of artists who’ve spent time in lookouts. Folks like Jack Kerouac, who spent a summer writing from a fire tower. 

Marina Avros

I like to be alone. I like nature. I like writing songs. I like creating. And I just … It just called to me. It’s hard sometimes. But I just feel this immense connection to the land. And I don’t know if I could get by anymore without a four-month break from society and civilization in the woods.

Arielle Duhaime-Ross

The isolation works for her. She actually does most of her songwriting during fire lookout season. Sometimes even in the tower itself. 

Marina Avros

When I’m up in the cupola, I’ll do some channeling exercises and just write lyrics and sort of just let it flow and then go downstairs and, you know, get out my guitar and try to piece it all together.

Arielle Duhaime-Ross

How important is this fire tower life to your music?

Marina Avros

It’s essential now. Maybe it’s a bit of a crutch. I have done some successful songwriting out of tower. But I really appreciate being able to say or sing whatever I want out loud and just have no fear of judgment, aside from my dog. 

Arielle Duhaime-Ross

Given the role fire towers play in her music, I wanted to know how she feels about all these new tools, like drones that are being used to detect fires, and if she worries about being replaced. Turns out, she doesn’t really.  

Marina Avros

I don’t think it’d be a good idea. We’re really cheap and super reliable so not a lot of opportunities for mechanical failure. We’re part of this network of fire detection. I’m a part of a bigger picture or a piece in the puzzle. If they took us away, they would be losing a really important factor in early fire detection.

Arielle Duhaime-Ross

That said, Marina sees the value of other technologies. 

Marina Avros

A drone would be able to check out spots I can’t see. We’re all in this together. 

Arielle Duhaime-Ross

I asked Marina if she’d written any songs about fire and whether she’d play one for me. She recorded this [one] at night, from her cabin, next to the fire tower. 

To hear Avros’s fire tower song, listen to the rest of the episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]> Delia Paunescu <![CDATA[The government keeps its use of facial recognition tech secret. The ACLU is suing.]]> https://www.vox.com/recode/2019/11/7/20953655/facial-recognition-technology-government-fbi-aclu-lawsuit-reset-podcast 2019-11-07T20:58:37-05:00 2019-11-07T17:00:00-05:00
A man orders coffee with the help of a face-recognition system at a stand of the AI Expo Africa in Cape Town. | Rodger Bosch/AFP/Getty Images

The ACLU is suing the FBI, the Department of Justice, and the DEA to get documents that explain how the US government is using facial recognition technology.

The lawsuit came after the government denied the ACLU’s request for information not only about how it is using this technology right now but also about how it plans to use that technology in the future.

“At the end of the day, when the people are in control, the government doesn’t have a right to make decisions about us without our input. But it’s hard to have input when you don’t know exactly what the government is planning to do,” said Jennifer Lynch, the surveillance litigation director at the Electronic Frontier Foundation (a nonprofit that focuses on civil liberties in the digital world).

On this episode of the Reset podcast, host Arielle Duhaime-Ross explores what we do know about how law enforcement and the government use facial recognition, as well as what’s at stake given what we don’t know. 

Today, nearly half of American adults are in facial recognition databases.

Russell Brandom, policy editor at The Verge, explains that the government first started with DMVs running facial recognition on all the photos in their system to see if any of them were the same person, “usually because it was Social Security fraud”:

They caught a lot of people this way. And now more recently on international flights, they’ve started doing it instead of fingerprints basically because it’s a simpler match. If you’re taking an international flight, you already have either a passport or a visa.

But it’s not quite so simple.

As Duhaime-Ross points out, big tech companies like Amazon and Microsoft have been selling facial recognition tech to various companies for business purposes while Amazon is also selling its facial recognition capabilities directly to law enforcement agencies, despite the fact that most citizens have never consented to their faces being used for these purposes.

That’s where the ACLU’s lawsuit comes in. Lynch says there’s currently a lot of secrecy surrounding how law enforcement is using facial recognition tech in the US and any additional information will help those who are trying to pass laws that’ll stop it from taking root in their communities:

A few years back, we got access to a PowerPoint presentation that the FBI gave. And in that PowerPoint presentation, they talked about sometime in the future wanting to be able to track people as they move from event to event. But we don’t know anything more about how far along they are in developing that kind of technology. So I think the ACLU’s lawsuit will get at information like that. Where is the FBI now or the Department of Justice or the DEA on implementing that kind of technology?

Listen to their entire discussion here. We’ve also shared a lightly edited transcript of ​Lynch’s conversation with Duhaime-Ross below.

You can subscribe to Reset on Apple Podcasts, Stitcher, or Spotify.

Jennifer Lynch

We haven’t seen face recognition come up too much in the court system yet. And I think that’s because it’s still somewhat novel technology in the United States. 

[But] we have talked to Congress about different kinds of bills that could be put forward to protect Americans from face recognition. We’ve been involved at the state and local level to pass laws that would ban face recognition or put a moratorium on the use of face recognition for a few years.

We also filed a couple lawsuits in the past to learn more about how the FBI was using face recognition. And those were Freedom of Information Act lawsuits, which are similar to what the ACLU has filed.

What we learned at the time was that the FBI had been doing these sort of backroom deals with the states to get access to their mugshot databases and build up the FBI’s own face recognition database. And all of that was without explaining what they were doing to the American public.

A few years back, we got access to a PowerPoint presentation that the FBI gave [in which] they talked about sometime in the future wanting to be able to track people as they move from event to event. But we don’t know anything more about how far along they are in developing that kind of technology.

So I think the ACLU’s lawsuit will get at information like that: Where is the FBI now or the Department of Justice or the DEA on implementing that kind of technology?

The other interesting thing about this ACLU lawsuit is that they aren’t limiting their request to face recognition. They are also interested in other forms of remote biometric identification technologies.

So, for example, gait recognition allows you to identify somebody based on how they walk. And if you were to combine gait recognition with face recognition, you could identify somebody who’s coming towards you and you can identify them as they’re walking away. We are starting to see these kinds of technologies used in concert with one another in China. And China’s selling their technology to other parts of the world, like Ecuador.

Arielle Duhaime-Ross

If I’m a law-abiding citizen, why should I be at all curious about how the US government is using this technology?

Jennifer Lynch

I don’t know that we can say that what the federal government is doing with face recognition is not going to affect us, even if we are law-abiding citizens.

Right now, the FBI has access to DMV photos in something like 20 to 30 states. I would argue that most of the people in those DMV files have not done anything wrong and yet they are being subjected to criminal searches without their knowledge [or] consent. And that could identify them as criminals or as suspects in a crime, because face recognition technology is not 100 percent accurate. So somebody could be identified for a crime they didn’t commit.

I hear this argument all the time that “I have nothing to hide so why should I care what the government is doing?” I think there are a couple other responses to that.

First, we need to all stand up for people who are constantly surveilled by law enforcement. I think we all know that communities face different kinds of surveillance depending on people, the color of people’s skin, depending on how people pray. And we all need to stand up for people like that, because people who are in those communities, they haven’t done anything wrong either but they are being subjected to excessive surveillance. 

I think the other argument I hear that somebody has nothing to hide. We all have something to hide. That’s the reason that we don’t let the government have cameras in our house. We need to have the ability to have privacy in our lives, to develop our thoughts and our beliefs. And it’s fundamental to living in a democratic society. We can’t really form our ideas about the government. We can’t think about how we feel in a society if the government is constantly watching us.

Arielle Duhaime-Ross

One thing that really struck me in what you just said was that you referred to the use of facial recognition technology as a form of search, law enforcement searching you, your body yourself. And I honestly don’t think that I personally have ever really thought about it in that specific way — that this is a search. Is that truly how we should think about this?

Jennifer Lynch

I think it is, depending on how it’s being used. So, for example, if law enforcement is trying to identify somebody from a photograph where they don’t know that person’s identity, who that person is, and they believe that that person has been involved in some sort of criminal activity, then they’re going to search a database to identify who that person is. And that is definitely a search. And it’s something that the Supreme Court has determined to be a search in other contexts.

Arielle Duhaime-Ross

What exactly is the FBI’s reasoning for not sharing information about how they use facial recognition technology?  

Jennifer Lynch

We haven’t seen that yet. So as part of this lawsuit, the FBI will have to file what’s called an answer, and they will do that 60 days after the ACLU has filed their complaint. And in that answer, the FBI might explain why they haven’t responded. But really what we’ll probably see over the next year or so is that the ACLU and the FBI will talk more about how the FBI is going to produce records in response to the ACLU lawsuit.

Arielle Duhaime-Ross

We’ve also heard that this technology isn’t the most accurate. What happens when law enforcement uses it to implicate someone of a crime that they’re not actually guilty of? How should we think about that?

Jennifer Lynch

In the United States, we have a rule that people are innocent until proven guilty. And if you are identified through a technology, that means that the state has decided that you are a suspect for a crime and it shifts the burden onto the person who is under suspicion to prove that they aren’t that person who committed the crime. That’s very different from how our criminal justice system is set up. The government should have to prove you’re guilty, not the other way around.

Arielle Duhaime-Ross

So it kind of sounds like if this is how things go down, this technology has the power to undermine the fundamental tenets of our justice system.

Jennifer Lynch

It does have [that] power … especially if there are no rules and especially if we don’t even know that it’s happening.

Arielle Duhaime-Ross

So what the ACLU is doing by asking these questions of the FBI, by trying to get more information on how the US government as a whole is using this technology, is that one way to prevent that future?

Jennifer Lynch

The more the public knows about how the government is planning to use face recognition technology and other biometric technologies, the better informed the public can be and the better able the public can be to make decisions about how the government should be allowed to use this kind of technology.

At the end of the day, when the people are in control, the government doesn’t have a right to make decisions about us without our input. But it’s hard to have input when you don’t know exactly what the government is planning to do. So that’s what the ACLU, whose lawsuit is trying to get at, to give us the information we need to make decisions and shape how our government should be.

Listen to the full episode and subscribe to Reset on Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.

]]>