Our entire financial system is built on trust. We can exchange otherwise worthless paper bills for fresh groceries, or swipe a piece of plastic for new clothes. But this trust—typically in a central government-backed bank—is changing. As our financial lives are rapidly digitized, the resulting data turns into fodder for AI. Companies like Apple, Facebook and Google see it as an opportunity to disrupt the entire experience of how people think about and engage with their money. But will we as consumers really get more control over our finances? In this first of a series on automation and our wallets, we explore a digital revolution in how we pay for things.
- Umar Farooq, CEO of Onyx by J.P. Morgan Chase
- Josh Woodward, Director of product management for Google Pay
- Ed McLaughlin, President of operations and technology for MasterCard
- Craig Vosburg, Chief product officer for MasterCard
This episode was produced by Anthony Green, with help from Jennifer Strong, Karen Hao, Will Douglas Heaven and Emma Cillekens. We’re edited by Michael Reilly. Special thanks to our events team for recording part of this episode at our AI conference, Emtech Digital.
Strong: For as long as people have needed things, we’ve… also needed a way to pay for them. From bartering and trading… to the invention of money… and eventually, credit cards… which these days we often use through apps on our phones.
Farooq: No one, 10 years ago, no one thought that, you know, you’d be just getting up from a dinner table and using Zelle or Venmo to send five bucks to your friend. And now you do.
Strong: The act of paying for something might seem simple. But trading paper for groceries…or swiping a piece of plastic for new clothes is built on a few powerful ideas that allow us to represent and exchange things of value.
Our entire financial system is built on this agreement… (and trust).
But this model is changing… and banks are no longer the only players in town.
[Sounds from an advertisement for Apple Card]
[Ad music fades in]
Announcer: This is Apple Card. A credit card created by Apple—not a bank. So it’s simple, transparent, and private. It works with Apple Pay. So buying something as easy as: *iPhone ding*.
Strong: It’s not just Apple. Many other tech giants are moving into our wallets… including Google… and Facebook…
[Sounds from Facebook’s developer conference]
Mark Zuckerberg: I believe it should be as easy to send money to someone as it is to send a photo.
Strong: Facebook Pay works through it’s social apps—including Instagram and Whatsapp—and executives hope those payments will one day be made with Facebook’s very own currency.
And beyond what we use to pay for things, how we pay for things is changing too.
[Sounds from an advertisement for Amazon One]
Announcer: Introducing Amazon One. A free service that lets you use your palm to quickly pay for things, gain access, earn rewards and more.
Strong: This product works by scanning the palm of your hand… and it’s not just for payments. It’s also being marketed as an ID. Something like this could one day be used to unlock the door at the office or to board a plane.
But letting companies use data from our bodies in this way raises all sorts of questions—especially if it mixes with other personal data.
Vosburg: We can see in great detail how people, for example, are interacting with their device. We can see the position in which they’re holding it. We can understand the way in which they’re typing. We can understand the pressure that’s being applied on the screen as people are hitting the keystrokes. All of these things can be useful with the combination of artificial intelligence to process the data to create sort of an interaction fingerprint.
Strong: I’m Jennifer Strong, and in this first of a series on automation and our wallets, we explore a digital revolution in how we pay for things.
Farooq: So, if you think about how we operate today, we primarily operate through central authorities.
Strong: Omar Farooq is the CEO of Onyx… from J.P. Morgan. It focuses on futuristic payment products.
Farooq: Frankly, the biggest central authority in some ways is, in the US, for the money purpose, is the US federal reserve and the U S Treasury. You pull out a dollar bill. It says U S Treasury. It’s issued by the, you know, in some ways, quote-unquote, the top of the house. The top of the house guarantees it. And you carry it around with you. But when you give it to someone, you’re ultimately trusting that central authority in how you are transacting.
Strong: This can be a good thing. The value of that otherwise worthless paper bill is guaranteed because it’s issued and backed by the US government. But it can also slow things down. And though we now take for granted being able to transfer money in real time, the ability to do so hasn’t been around that long.
Farooq: Payments actually do, as a technology, evolve somewhat slowly. Just to give you an example, the U S recently, a couple of years back, launched the real-time payments scheme, which literally was the first new payments, you know, sort of, rails in the U S for decades. As crazy as that sounds.
Strong: A payment rail is the infrastructure that lets money move from one place to another. And those “real time payments” are a big deal because until recently when money left your account it took time, often days, before it reached its destination.
It’s why we can send money through apps like Venmo and hear the ding that it’s been received on the other person’s phone just a few seconds later. Also, Venmo’s chief competitor, called Zelle, only exists because of unprecedented cooperation between otherwise competing banks.
Farooq: I think where the world is going is towards more open platforms where it’s not just one party’s capabilities, but multiple parties’ capabilities that come together. And the value that is generated is by the ability for anyone to connect to anyone else. So I think what we are seeing is a rapid evolution in the digital sphere where more and more payment types, whether they are wholesale or retail are going into new modes, new rails, 24/7, 365, the ability to pay anyone anywhere in any currency. All those things are basically getting accelerated.
Strong: This is where cryptocurrencies could come in. Which isn’t just about digital money.
Farooq: We believe that there’s a path forward where money can be smarter itself. So you can actually program the coin and it can control who it goes to.
Strong: In other words, the trust we usually place in banks or governments would be transferred to an algorithm and a shared ledger.
Farooq: So you’re almost relying on that decentralized nature of the algorithm and say, “I think I can trust your token coming to me” because there’s, you know, X… X thousand or X hundred-thousand copies of a ledger that shows you as the owner of that token. And then when you give it to me, All those copies get updated. And now this shows me as the owner of that token.
Strong: And not only could this make payments faster and more seamless. It could also help people who’ve been largely excluded from the banking system.
Farooq: No matter what we do, we cannot really get around this Know Your Customer issue. And I think, you know, our view is that the tech is almost there, but the regulation and the infrastructure around it is not there yet. But, what we do want to do is we want to create these decentralized systems where these people can, over time, be included.
Strong: But sorting out the tech… is just one side of the coin. There’s also a need for better regulation.
Farooq: But I think it’s unfortunately a little bit more than what a bank could do. I think this.. some of these things rise to the level of like, you know, how does a government, or how does a state really enable identity at a global level? And I think that’s why when you look at China or you look at Nordics or some of those countries, I mean, you have national IDs and you have a very standardized method of knowing who someone is.
Strong: And the shift it allows in banking can be transformative…
Farooq: So if you look at a country like India, India has made dramatic progress in how many people have gone from being unbanked to banked in terms of having a wallet on their mobile phone. So I think these technologies are going to turbocharge people’s ability to come into this ecosystem. What I would hope as someone who grew up in the developing world before migrating here is that you would make those connections so, you know, everyone in those countries has access to markets—to bigger markets. So I mean, whether you’re sitting in Sub-Saharan Africa or you’re sitting in like, you know, a village in India or Pakistan or Bangladesh, wherever, you can actually sell something through Amazon and get paid for it. I mean, you know, those sorts of things. I think there’s tremendous potential human potential that could be unlocked if we could take payments in a digital manner to some of those parts of the world.
Strong: And this vision?… extends not only to connecting anyone, anywhere to a bank… but also anything with an internet connection.
Farooq: doing some initial R and D work in the IOT space, which is, if, you know, I mean, if one day your fridge had to order milk by itself. Like, does it have to go through your bank or could it just send the money to someone who’ll deliver your milk?
McLaughlin: Every device you use has potential to be a commerce device and our network brings that together.
Strong: Ed McLaughlin is president of operations and technology for MasterCard. He’s speaking at our A-I conference, EmTech Digital.
McLaughlin: So, what all of that connectivity results in? Is.. bringing together pretty much every financial institution in the world, tens of millions of merchants, governments, tech CO’s, and all of that, which results in billions of transactions a year we see. MasterCard across all of those devices and cards is serving about two and a half billion accounts. So we get the data and transactions from a Facebook sized population, if you think about that… And as far as the scope goes, we’ve been probably seeing 20 to 25% of all internet transactions outside of China—since there was an internet.
Strong: But this connectivity creates its own set of new problems. Maybe you’ve had the experience of going out of town and suddenly your card stops working because the change of location triggered a fraud alert.
McLaughlin: One of the keys in applying AI is how you frame the question and our teams very early on and said it wasn’t to stop transactions. It was to make sure as many good transactions as possible made it through.
Strong: Another key is to have an abundance of data.
McLaughlin: It’s a massive in-memory grid in our network that holds over 2 billion card profiles with about 200 analytical vectors on it. And we make decisions in every transaction that flows through. We have less than 50 milliseconds to make that decision. So in order to do that, we have 13 different AI technologies that we’ve modeled and experimented over the years that we apply to it.
Strong: Banks are also turning to A-I to look for money laundering. In the physical world, organized crime is often hidden behind the storefronts of real businesses. And in the digital world? Hiding is even easier.
Illegal money can quickly change hands dozens of times and cross borders until there’s no clear trail back to its source. It’s a massive problem. And most of it goes undetected. It’s possible only one percent of the profits earned by criminals gets caught. And the turmoil of the global economy over the last year has only made things worse.
McLaughlin: Our adversary.. They’re using AI too. And if you look online, it’s just bots fighting bots. So you have to pick up things you weren’t looking for before, like low and slow attacks where they stay inside, what looks like acceptable tolerances, but they’re constantly probing or doing a tumbler attack on your systems. Hard to pick up. When COVID hit, you know, the world moved online. Spending patterns shifted dramatically. And what we were able to do because the AI’s are rich enough and look at so many different variables.. We were able to really tell you’re still you and you’re just behaving a little bit differently.
Strong: And the types of attacks change too…
McLaughlin: So we saw one attack factor, which was pretty amazing is they thought, okay, people won’t block transactions for personal protective gear. It’s a specific merchant class we have. And we saw the fraudsters pile on in trying to get transactions through because they figured nobody would be blocking. The good news is we look at enough other elements that we could immediately pick that up and block those transactions.
Strong: They’re building machine learning tools to identify patterns of normal activity. And to flag outliers when they’re detected. Humans can then double check those alerts and approve or reject them.
McLaughlin: We constantly have AIs running also, not just blocking the fraud or looking at it, but I’m just calling it weirdness detection—where we’re constantly predicting what we would expect to see. In fact it’s a great way to step into AI because you have KPIs you’re already tracking. Try to start predicting them. When you see something which is an immediate deviation from it, the first thing we actually do is say, what’s going on here? So we may see something the model hasn’t caught up to, we just throw a rule to block it. And we can do that instantly.
Strong: The payments industry used to be slow moving… but it’s adapting to a world where any device might one day be connected to a payments network… including self driving cars.
McLaughlin: So whether you’re using your browser to order online, if it’s your iPhone, we’re using an Apple Pay to tap, or Mercedes just announced that, uh, they’re going to be connecting their cars to gas pumps. So you can simply drive up and authorize your transaction, right from your car. And in fact, as things move away from the card and to devices, we’re seeing even more data coming in through the network.
Strong: We’ll be back… right after this.
Strong: With more and more of our financial lives being documented, tracked and mediated online, that data turns into fodder for AI—which is being enlisted into a whole host of other roles with payments.
Woodward: People have a really complex relationship with their money. It can be stressful. It’s often boring a lot of the time.
Strong: Josh Woodward leads the Google Pay team for the US. He sees it as an opportunity to change not just payments…but the entire experience of how people think about…and engage with…their money.
Woodward: And so what we’re trying to do as a team is think about how can we simplify that relationship with money where people feel in control and they feel confidence when they’re using our app and seeing how their spending is going in and out.
Strong: Google Pay began as a peer to peer payment solution—where the main goal was digitizing the plastic cards in your wallet. But over the years, it’s evolved into a tool meant to help you more holistically manage your finances, and relationships with businesses.
Strong: And it’s taken some cues from social media. Instead of card numbers or accounts, transactions are organized around pictures of people and businesses you’ve recently paid.
Woodward: We realized that transactions, in some ways, the.. the money, that the digits, the dollars and cents, is secondary. It’s a lot more about the person or the memory around that transaction. So we’ve tried to bring that out. Similarly, we’ve taken that same relationship based design and applied it to businesses. And this is something that’s very different. So when you look today at our home screen, // what you see is actually the icon of the business. And when you tap on that, you are taken to that business page where you can actually. Really see, like your relationship with the business. If you have a loyalty card you can see that there, you can see how your points are progressing. So the next time you go buy, you can get 20% off for example. And so we’ve tried to create this… Really almost like a threaded relationship of all your activity with that business inside the Google Pay app a little bit like Gmail, threaded email messages.
Strong: It also lets users sort transactions in a way that mirrors a web search.
Woodward: So you can do things like search for food. And you’ll get all of the transactions at places where you bought food and Google Pay can understand that this restaurant, for example, is a restaurant. You don’t have to go in and manually categorize that. Or you can get more specific and do things like a search for Mexican restaurants. And it’ll just take that subset of Mexican restaurants. There’s no part of that transaction that has the phrase, Mexican restaurant in it. Google Pay’s able to make that connection for you.
Strong: And using computer vision…it can sort through photos of receipts.
Woodward: What we’ve been able to do in Google Pay, again with someone’s permission, this feature is off by default, is that you can say, I want all the photos I’ve taken of receipts to be searchable in Google Pay. And what that allows you to do is actually search very specifically for individual items that are printed on the receipt. So for example, a couple of months ago, before Christmas, I bought a shirt, uh, it was a Christmas present from Lulu. I can go into Google Pay now and search for “shirt.” And that Lulu receipt comes up.
Strong: It’s designed to give users a greater sense of control over their spending.
Woodward: It creates a place where you get that full picture. And that’s what we’ve seen. Time and time again, in the research and in talking to people is that different apps have provided different slices of that picture, but being able to bring it all together is really what we aspire to.
Strong: It’s one more way our lives might become a little easier and more efficient with the help of technology… But also where the gathering… filtering… and processing… of vast amounts of personal data raises big questions… even before we get to things like paying with our faces or gestures… or how all of that data… might mix with the rest of our massive data trails.
And longer-term, what would it mean for companies like Facebook to establish their own currencies and take over the global payments system?
It’s worth asking whether we as consumers really get more control over our finances… or companies get more control over us…
Bennett: We couldn’t have imagined something like Siri or Alexa. You know we just thought we were doing just generic phone voice messaging… and so in 2011 when suddenly Siri appeared, it’s like, “I’m WHO??” [laughing]… “WHAT??”…
Strong: We look at what it takes to make a voice… and how that’s rapidly changing.
Strong: This episode was produced by Anthony Green, with help from Jennifer Strong, Karen Hao, Will Douglas Heaven and Emma Cillekens. We’re edited by Michael Reilly. Special thanks to our events team for recording part of this episode at our AI conference: Emtech Digital.
The Download: metaverse fashion, and looser covid rules in China
Fashion creator Jenni Svoboda is designing a beanie with a melted cupcake top, sprinkles, and doughnuts for ears. But this outlandish accessory isn’t destined for the physical world—Svoboda is designing for the metaverse. She’s working in a burgeoning, if bizarre, new niche: fashion stylists who create or curate outfits for people in virtual spaces.
Metaverse stylists are increasingly sought-after as frequent users seek help dressing their avatars—often in experimental, wildly creative looks that defy personal expectations, societal standards, and sometimes even physics.
Stylists like Svoboda are among those shaping the metaverse fashion industry, which is already generating hundreds of millions of dollars. But while, to the casual observer, it can seem outlandish and even obscene to spend so much money on virtual clothes, there are deeper, more personal, reasons why people are hiring professionals to curate their virtual outfits. Read the full story.
Making sense of the changes to China’s zero-covid policy
On December 1, 2019, the first known covid-19 patient started showing symptoms in Wuhan. Three years later, China is the last country in the world holding on to strict pandemic control restrictions. However, after days of intense protests that shocked the world, it looks as if things could finally change.
Beijing has just announced wide-ranging relaxations of its zero covid policy, including allowing people to quarantine at home instead of in special facilities for the first time.
Uber’s facial recognition is locking Indian drivers out of their accounts
Uber checks that a driver’s face matches what the company has on file through a program called “Real-Time ID Check.” It was rolled out in the US in 2016, in India in 2017, and then in other markets. “This prevents fraud and protects drivers’ accounts from being compromised. It also protects riders by building another layer of accountability into the app to ensure the right person is behind the wheel,” Joe Sullivan, Uber’s chief security officer, said in a statement in 2017.
But the company’s driver verification procedures are far from seamless. Adnan Taqi, an Uber driver in Mumbai, ran into trouble with it when the app prompted him to take a selfie around dusk. He was locked out for 48 hours, a big dent in his work schedule—he says he drives 18 hours straight, sometimes as much as 24 hours, to be able to make a living. Days later, he took a selfie that locked him out of his account again, this time for a whole week. That time, Taqi suspects, it came down to hair: “I hadn’t shaved for a few days and my hair had also grown out a bit,” he says.
More than a dozen drivers interviewed for this story detailed instances of having to find better lighting to avoid being locked out of their Uber accounts. “Whenever Uber asks for a selfie in the evenings or at night, I’ve had to pull over and go under a streetlight to click a clear picture—otherwise there are chances of getting rejected,” said Santosh Kumar, an Uber driver from Hyderabad.
Others have struggled with scratches on their cameras and low-budget smartphones. The problem isn’t unique to Uber. Drivers with Ola, which is backed by SoftBank, face similar issues.
Some of these struggles can be explained by natural limitations in face recognition technology. The software starts by converting your face into a set of points, explains Jernej Kavka, an independent technology consultant with access to Microsoft’s Face API, which is what Uber uses to power Real-Time ID Check.
“With excessive facial hair, the points change and it may not recognize where the chin is,” Kavka says. The same thing happens when there is low lighting or the phone’s camera doesn’t have a good contrast. “This makes it difficult for the computer to detect edges,” he explains.
But the software may be especially brittle in India. In December 2021, tech policy researchers Smriti Parsheera (a fellow with the CyberBRICS project) and Gaurav Jain (an economist with the International Finance Corporation) posted a preprint paper that audited four commercial facial processing tools—Amazon’s Rekognition, Microsoft Azure’s Face, Face++, and FaceX—for their performance on Indian faces. When the software was applied to a database of 32,184 election candidates, Microsoft’s Face failed to even detect the presence of a face in more than 1,000 images, throwing an error rate of more than 3%—the worst among the four.
It could be that the Uber app is failing drivers because its software was not trained on a diverse range of Indian faces, Parsheera says. But she says there may be other issues at play as well. “There could be a number of other contributing factors like lighting, angle, effects of aging, etc.,” she explained in writing. “But the lack of transparency surrounding the use of such systems makes it hard to provide a more concrete explanation.”
The Download: Uber’s flawed facial recognition, and police drones
One evening in February last year, a 23-year-old Uber driver named Niradi Srikanth was getting ready to start another shift, ferrying passengers around the south Indian city of Hyderabad. He pointed the phone at his face to take a selfie to verify his identity. The process usually worked seamlessly. But this time he was unable to log in.
Srikanth suspected it was because he had recently shaved his head. After further attempts to log in were rejected, Uber informed him that his account had been blocked. He is not alone. In a survey conducted by MIT Technology Review of 150 Uber drivers in the country, almost half had been either temporarily or permanently locked out of their accounts because of problems with their selfie.
Hundreds of thousands of India’s gig economy workers are at the mercy of facial recognition technology, with few legal, policy or regulatory protections. For workers like Srikanth, getting blocked from or kicked off a platform can have devastating consequences. Read the full story.
I met a police drone in VR—and hated it
Police departments across the world are embracing drones, deploying them for everything from surveillance and intelligence gathering to even chasing criminals. Yet none of them seem to be trying to find out how encounters with drones leave people feeling—or whether the technology will help or hinder policing work.
A team from University College London and the London School of Economics is filling in the gaps, studying how people react when meeting police drones in virtual reality, and whether they come away feeling more or less trusting of the police.
MIT Technology Review’s Melissa Heikkilä came away from her encounter with a VR police drone feeling unnerved. If others feel the same way, the big question is whether these drones are effective tools for policing in the first place. Read the full story.
Melissa’s story is from The Algorithm, her weekly newsletter covering AI and its effects on society. Sign up to receive it in your inbox every Monday.