Skip to content

Artificial Intelligence Podcast: Experts Discuss Critical Trends

Artificial intelligence will drive a $15.7 trillion-dollar economy by 2030. To put this in perspective, the top 5 technology companies today have a combined value of $4 trillion, which includes Apple, Amazon, Microsoft, Google and Facebook. The annual global technology spend is similar is $3 trillion. Over the next decade AI will drive a market 5x the size of tech’s current global spend and 4x the size of five of the world’s most valuable companies combined. Industries that AI is already transforming include tech, health care, finance, automotive and more.

In episode 4 of Tech Lightning Rounds, you’ll hear a 360-degree view on AI trends from experts who are on the cutting edge of developing and leveraging AI applications. Mastercard provides insight into how AI and natural language processing (NLP) is helping to combat fraud. Ann Cairns, Executive Vice Chairman of Mastercard, discusses how Mastercard knows if someone is using your phone to access your bank accounts and what kind of signals NLP looks for when detecting money laundering.

Beth Kindig of Intertrust also interviews Chistophe Coutelle of Element AI, who discusses how AI is being used in capital markets to inform portfolio managers and stock traders. Coutelle also discusses how AI has been used to evaluate cyber bullying on sites like Twitter.

The last interview is with Dragana Krcum of Visage Technologies on the use of biometrics, such as facial recognition, eye tracking, iris detection, head tracking and palm vein detection. This interview is especially candid as a biometrics expert also discusses the challenges we face with privacy globally, including China, and why the General Data Privacy Regulations in Europe may have come at the right time.

Please Subscribe and Leave a Review for the Podcast Here

Apple PodcastGoogle PodcastSpotifyStitcher

TRANSCRIPT:

00:24 Beth Kindig: Welcome to Tech Lightning Rounds. I’m your host, Beth Kindig. This podcast interviews key people with deep expertise on one topic for a 360-degree view. One difference between this podcast and the other podcasts you listen to is that I hold short interviews, called Lightning Rounds, with the goal of getting you a lot of compelling information very quickly so you can get on with your day.

00:51 BK: In these lightning rounds, I spoke with a diverse range of experts in artificial intelligence. We discuss how AI combats fraud, how AI informs capital markets, and we also discuss tough issues like privacy, especially as it relates to facial recognition. You’ll hear from Mastercard, on how the company uses behavioral biometrics and natural language processing.

01:14 Ann Cairns: Where we see things moving erratically and suspiciously and being split and recombined, and so on, in a very fast pace. We can actually track that and can light it up on a dashboard as if it was a cell diagram, and you actually were looking at, say, the spread of a virus.

01:33 BK: From Element AI on various uses for AI, including predicting the stock market and preventing online bullying.

01:39 Christophe Coutelle: Another one is a product called the Trade Scheduler, which is about identifying the right moment for portfolio managers to sell or buy specific stocks.

01:48 BK: And definitely don’t miss the last lightning round with Visage from Sweden, where the interview gets serious about facial recognition and how privacy addresses a future where citizens are being tracked by face, iris detection, head movements, and palm veins.

02:03 Dragana Krcum: So, head tracking is tracking of the head movements and head position and rotation, and so on. But in addition to that, we also track the facial feature points of the user. For example, we track 75 points.

02:18 BK: In my first lightning round, I speak with Ann Cairns, the vice chairman of Mastercard, who discusses how the patterns and movements of how you touch your phone as an individual, which are unique to you, can help determine if you’ve been hacked.

02:33 BK: Let’s talk a little bit about artificial intelligence and how Mastercard uses artificial intelligence to prevent fraud. Can you give us some background, how much fraud is occurring in the world, and what has AI done to combat that?

02:49 AC: Well, in the financial world, fraud is usually in the physical world at a sort of single-digit percentage as you’re moving money around the world. And as you go on to the internet and you’re using it in that way, then the numbers start creeping up. And so what all of the big players are doing now are using new fraud tools, many enhanced by artificial intelligence, to try and reduce the level of fraud when you’re working online or in the digital space. And the sort of things that are happening now are, for example, behavioral artificial intelligence. We have a telephone in front of us, if your phone can make payments, and I certainly picked it up and try to make a payment using your phone, it would know from the key strokes that I’m using if you do some of our technology, it would know that it isn’t you, it’s me that’s doing that. Or the phone and the way that we use it differs from person to person. That’s the kind of artificial intelligence we have now.

03:56 BK: Is it the speed of the touch, or what’s different about how I touch a phone versus how you touch a phone or perhaps a hacker touches a phone?

04:04 AC: Yeah. It’s not one simple thing. Apparently, I’m older than you, and so I probably hold the phone further from me and I probably hold it flatter. You’re younger, and you would hold it up towards your face because I’m long-sighted. And also you might be very faster at keying with your fingers or your thumbs. I may just be a one-finger keyer, that I didn’t learn to type when I was a kid. It’s all sorts of things like this that are combined.

04:39 BK: How can AI help prevent money laundering? Can you expand on natural language processing?

04:46 AC: I’ve mentioned before that we bought VocaLink and one of the things, when people are laundering money, is they tend to move it very quickly from bank account to bank account. And maybe what they do is move it back and forward between a number of different bank accounts several times, and sometimes they split it so that it actually… Some pieces move to one bank account, another piece moves to another bank account, and they create this highly confusing pattern over everything that they’re doing. And now, we have the ability, through artificial intelligence, when banks have hooked up to us, to actually look at those patterns and identify that pattern of money movement and where we see things moving erratically and suspiciously and being split and recombined, and so on, in a very fast pace. We can actually track that and can light it up on a dashboard as if it was a cell diagram, and you actually were looking at, say, the spread of a virus, for example. It’s quite fascinating to see this.

05:54 BK: Mastercard is one company that has an advantage in helping entrepreneurs and small business owners, who live in remote areas, access bank accounts. Ann had been on a panel earlier that day, and I asked her to expand on why it’s important for people, who want to succeed, to have a credit line.

06:10 BK: Earlier this morning, at MWC, you had talked about the fact that 1.2 billion additional people now have access to a bank account, one in five bank accounts are not in use, and approximately two-thirds of mobile money accounts are dormant. Why is that important and what will Mastercard do to change that?

06:31 AC: We’re beginning to see patterns. As you said, there’s a lot of dormant accounts out there, and yet it’s so needed the ability to be able to transact electronically because, once you start doing that, you start to do things like build up a credit history. And often around the world, particularly if you’re a woman and you don’t have any access to credit, having something that showed you regularly paid your rent, that the shop you were running was actually performing well, and your business was flowing well, all of these things enhance your business and allow you to be able to get more access to finance and to growth.

07:15 BK: So, Mastercard is doing certain things to create a more inclusive environment then?

07:22 AC: Definitely we’re creating a more inclusive environment. And some of the ways we’re doing that is actually going into partnerships with companies, like Unilever, where we can use their supply chain information to be able to reach small shopkeepers and get those people credit, because we can put in the electronic payment infrastructure to look at it end to end, share the data with banks, and show the banks that actually the small shopkeeper is running the business well and that that bank should lend that shopkeeper money. We’ve done that in certain parts of Africa, and when we launched that, we found that instead of waiting for the shelves to get empty, so the shopkeeper had cashed buy more products from Unilever, the bank could lend the money and the shopkeeper could order ahead, and the sales grew 20%. It’s a win-win-win situation, because the bank is happy, Unilever is happy, the shopkeeper is happy. We’re happy, because we’re capturing digital transactions.

[music]

08:30 BK: Capital markets is an especially intriguing industry that artificial intelligence will affect, as the financial upside the better predictions is nearly limitless for stock trades and other investments. In this lightning round, you’ll hear from Christophe Coutelle, from Element AI in Montreal, dive deep into the products that are currently available for investors.

08:52 BK: How does AI inform capital markets? Can you give me a few use cases there?

08:56 CC: AI is very good at getting insights from a lot of data, so that’s the example I give to you with the Port of Montreal. But in capital markets, we’re really able to identify trends, insights, signals from a lot of data that is coming from a lot of different systems. AI in capital markets is really good at not only getting those insights but also predicting and, again, based on what happened in the past. For instance, we have developed and we’re currently rolling out two products with a large financial institution in Asia. Those two products, one is called the Portfolio Rebalancer. This is really about identifying the stocks that portfolio managers should be focusing on based on historical data. Another one is a product called the Trade Scheduler, which is about identifying the right moment for portfolio managers to sell or buy specific stocks. Again, this is based on a lot of different insights that AI is able to extract, and that’s where we believe AI has an important role to play in capital markets.

10:05 BK: In that situation, would those products be available to individual investors or hedge funds, mutual funds? Who would that be available to?

10:13 CC: Currently, we’re targeting more large enterprises, because AI is very data hungry, so we need to train our models with data and that’s why we’re targeting large companies at the beginning. But once we’ll train those models with these data, we’re gonna be able to target smaller and smaller companies, and eventually perhaps go down to individual investors that were really not there yet.

10:40 BK: Going back to the AI for capital markets, if the AI is available to the bigger companies, is that gonna create more of a lopsided economy where the top couple percent are gonna have an even bigger advantage if they’re able to access AI before the rest of us?

11:00 CC: Okay. That’s a fair question, because they are getting the benefit from the products as they have the data. But again, our objective is really, once the models are trained to make those products available for Tier 2, Tier 3, and then smaller customers, even eventually individual investors, that’s something that can happen very quickly. I think the entire advantage wouldn’t last very long in any case.

11:28 BK: Because Element AI is involved in many different industries, such as manufacturing, retail, and insurance, I asked Christophe to expand on the good AI will do for society, and he gives some specific examples, including how to improve employment opportunities and how to address cyberbullying.

11:45 BK: Where do you see AI making the biggest impact? What’s the biggest problem that AI will solve?

11:51 CC: I think the biggest problem that AI will solve will definitely getting rid of what we call the routine work. It’s not the end of work, it’s the end of routine. I think that’s really where AI will have the biggest impact in, again, one, being able to identify insights and signals from a lot of data. Then from those insights, being able to do more accurate predictions. And then with those predictions and this planning, being able for the AI products to execute on them, perhaps making recommendation or even going down to automation and some of the very specific low value tasks, and then freeing up some time for the people to focus on the higher value tasks.

12:39 BK: What is AI for Good? Can you tell us a story there?

12:42 CC: Okay. AI for Good is something that is really aligned with the values of Element AI. We have an office in London that is dedicated to AI for Good. An example of that is perhaps an announcement that you’ve seen here back in December. It’s a project that we’ve done together with Amnesty International, which was aiming at identifying potentially offensive tweets on social media. And so we train our model, and we did that with a sample of 1000 female politicians and journalists, and we’ve identified that there was a very significant amount of offensive tweets for those populations. So, we’ve been extrapolating that model up to millions of tweets, and that project was actually published by Amnesty International. What we did was to release a data set that people can now use to trend their own models and identify potential offensive tweets on social media. That’s an example of what AI for Good can do. We have other projects in the pipe at the moment with that office in London, but there will be announcements probably later on this year.

14:09 BK: In the example, your model found that there were basically offensive tweets, more likely to be offensive tweets towards women than men on Twitter. What does Twitter do with that information after you provide it to them, a company like Twitter?

14:27 CC: A company like Twitter actually reached out to Amnesty International. Amnesty was the one owning the project. We’re just the technical enabler behind them. I know they had followup meetings based on that, and they have a very positive reaction in a sense that they’ve been trying to understand how the study had been made and how the project was handled so that they could actually improve their capacity to tackle that problem. Again, we don’t own the relationship there, we’re just behind and providing support to Amnesty International. But I know that those social media, and especially Twitter, they have reacted very positively to that and looking for solutions.

[music]

15:16 BK: The next lightning round has really lingered with me for some time, as Dragana Krcum, of Visage from Sweden, talks intelligently and frankly about important issues around not only facial recognition, but also iris tracking and palm vein detection. Listen as she dissects what you need to know about biometrics.

15:34 BK: Where are we at now with biometric authentication, and where do you see biometrics going in the next five to 10 years?

15:41 DK: There are really a lot of use cases for biometrics and verification. Most popular ones are face tracking, iris detection and tracking, fingerprint, and voice recognition. They all have pros and cons basically. We choose to be in the face domain because that seemed closest to us, as this is how we started as a company. Nowadays, there are many, many more use cases like palm vein detection and behavioral biometrics.

16:22 BK: Earlier you had mentioned iris tracking, are you talking about eye tracking? What do you mean by that?

16:29 DK: There’s a difference. You have eye tracking, which tracks the outer corners of the eye, and then there’s iris tracking which is a section in the eye. The difference is that, if we’re gonna talk about the purposes for each application, eye tracking is much easier because iris tracking and detection and recognition, it is very accurate, but it’s very expensive to get machines, hardware that can perform good iris tracking.

16:58 BK: Why would anyone need to track an iris?

17:02 DK: Because, like fingerprint, iris pattern for each person is unique. So, this gives it… It’s a great possibility to determine the identity of a person.

17:13 BK: We can’t just stick with fingerprints, I guess?

17:16 DK: Fingerprints is something that’s been around for a while now, but, of course, we’re looking for alternatives because, I think, fingerprints are… Head tracking and facial recognition is less invasive in terms of user experience and processing, because if you have to have access control that uses fingerprints, you have to have a fingerprint reader. But, for example, for face recognition, all you need is a camera and a software. The person doesn’t really need to be involved in the process of the authentication.

17:50 BK: You had mentioned potentially using this technology at borders, is that currently happening or is that more in the future?

17:57 DK: Actually, it’s currently happening already. It’s really big in Europe. You take a photo of yourself on the border, and then it is automatically compared to your passport photo, which is how we do facial recognition at the moment. But, of course, there are many other uses for facial recognition that you can use for security purposes. And I believe that this type of user identification will be even more in use in the years to come, especially at border control.

18:33 BK: And when you mention head tracking, you’re talking about the movements of the head or the positioning of the head. How would my movements differ from yours or someone else’s?

18:44 DK: Head tracking is a tracking of the head movements and head position and rotation, and so on, but in addition to that, we, at Visage and other companies, also track the facial feature points of the user. The number of areas for each company, but, for example, we track 75 points, and the combination of these points, in addition to the head position tracking, gives you outputs on the user’s face.

19:11 BK: Because Dragana is so knowledgeable on the space, I prompted a few questions that I thought were important to discuss around privacy. With Visage being from Sweden, we go into depth on how Europe has set a good example globally on how to address privacy issues.

19:26 BK: A lot of my listeners are probably wondering about China’s use of biometrics. To most standards today, it might feel fairly invasive to have facial recognition on street corners, to access airline flights. From a technology standpoint, like someone who really works with this technology on a daily basis and talks about it, do you think this is invasive, and why or why not?

19:48 DK: For this use case, in China, I think there is pros and cons to each situation. The pros, of course, would be that… In China, I think it’s good that the legislation for this facial recognition and social point system, it’s governed by the state, and it has laws and regulations on how it operates. That is a good thing. However, I don’t think it’s a good thing to invade so much on the privacy and everyday lives of the citizens so they would live in fear of repercussions for their deeds. I am totally up for all new technologies that will try to minimize fraud in business and other segments or help the law enforcement, so in that sense, I think it’s a good thing, even though it has its flaws.

20:43 BK: Where do you see us with biometrics, basically, in 10 years from now?

20:50 DK: Do you mean in Europe or in general?

20:53 BK: Let’s say Europe and North America.

20:54 DK: Yeah. Not sure if you know, but since last year, we actually introduced a new regulation for privacy. It’s called GDPR. Things are moving in that direction in order to protect the privacy of the users and their data. However, I think that biometrics is still gonna be strong and grow even larger because we will just have to make up some new laws that will protect the user but also introduce these technologies to the market.

21:31 BK: What is the proper use of biometrics? What’s the middle ground between what the general population wants and needs, and what companies or governments would want or need? What’s the compromise?

21:42 DK: There’s the benefit of our products that some other companies do not have. One of those benefits would be that we work on all major platforms, and we are completely GDPR-compliant since we do not store data by default.

22:01 BK: And you’re from Sweden, is there any reason why… Is there anything special about Sweden and why maybe this technology is coming out of Sweden? I think it’s great you are mentioning the GDPR. I’ve written extensively about the GDPR, and maybe this kind of thing does need to come from Europe because it’s really pushing hard for consent and the ability to export and delete your data. Can you give me some background on Sweden, and Europe in general, as to why maybe Visage Technologies being from those regions is important?

22:34 DK: Visage is a Swedish company. Its headquarters are in Sweden, but we also have an office in Zagreb, in Croatia, which is also in European Union. The combination of these two factors, I think, both of us being European, is something that helps a lot with our view on the market and in which way we want to keep innovating, because we do have these regulations that we want to follow, so I think it’s a good steering point for us as a technology company.

[music]

23:11 BK: Thank you for listening to the artificial intelligence episode of Tech Lightning Rounds. Please subscribe and leave a review in iTunes to support the production of this podcast.

23:21 S1: This episode was brought to you by Intertrust Technologies and Modulus, helping you build data-driven businesses. Go to intertrust.com for more information.


Sign up for Analysis on the Best Tech Stocks

Beth KindigI’m an industry insider who writes free in-depth analysis on public tech companies. In the last 12 months, I predicted Facebook’s Q2 crash, Roku’s meteoric rise, Uber’s IPO flop, Zoom’s IPO success, Google’s revenue miss and more. Be industry-specific. Know more than the broader markets. Sign up now. I look forward to staying connected.

If you are a more serious investor, we have a premium service that offers institutional-level research and entry/exit options. This membership offers a competitive edge in identifying growth opportunities and reducing risk in the tech sector. Learn more here.

Join 3,003 other tech investors who receive weekly stock tips:




    Published inArtificial IntelligenceTech Podcast

    One Comment

    1. Thanks for sharing this post,
      is very helpful article.

    Leave a Reply to simon Cancel reply

    Your email address will not be published. Required fields are marked *