Live from Money 20/20 Europe: AI in Fintech with author Angelique Schouten

Speaker A: Foreigners have played with ChatGPT, prompting it on a wide range of subjects concerning fintech, finance, tech and even our personal lives. In fact, a few weeks ago I asked you to write a love letter to Jan, the guy I'm going to marry in a few weeks. But I decided to do it myself. Um, anyway. But none of us have actually engaged with ChatGPT as a full time partner on a single project. Einstein Schroute. Our guessing the episode of Moniepod. Has she wrote the book Rising AI Tech Demystified with ChatGPT? Not in months, but in one long and very intense week. We talked to her today about the process and how it was to treat ChatGPT. As a real life co author, she dives into the lessons learned. How to go from copywriting to prompt writing. Has it changed her mind on the power of AI? Can one actually combine human creativity and AI? And of course, would she ever do it again?

Speaker B: Give me that, uh, old, um, fashion morphine. Give me that old fashioned morphine. Give me that old fashion morphine. That's, uh, good enough for me.

Speaker C: Welcome, Angelique, live from Money 2020. 2023. Welcome Angelique. You're looking wonderful today, by the way. Angelique.

Speaker B: Thank you. You too.

Speaker C: We're going to talk about AI and it's a hot topic at, uh, even this morning, a big article in the Guardian about Sam Altman and Geoffrey Hinton basically warning AI as a threat like nuclear war and pandemic. So we're going to dive into that. But first, uh, let's Talk about Money 2020. You were here yesterday probably, yes. Something particular that caught your eye?

Speaker B: Yeah, for sure. I think one of the, uh, presenters had a book published which I really liked and the book title is actually Fall in Love with the Problem and Not the Solution, which I think a lot of companies miss. They're so focused on what they're doing their own bit and don't talk to the customer and to the end user. So I really, really particularly like that, uh, phrase, but also the mindset and mentality and I hope that a lot of people here will adopt it.

Speaker C: Oh, great. So we have to buy that book for sure. All right.

Speaker A: Well, as a creative maverick and new tech fanatic, Angelique spent a decade defying convention with Fintech Open, the world's first cloud native core banking platform, the bank to Bank Skyward, boasting over 100 billion assets under administration after its startup phase and many more startup and scale initiatives followed. She co authored bestsellers like Wiley's Wealth Tech book, Forbes Monkey money mind investigating the psychology behind money and sharing her journey from 100,000 student debt to savvy investments, whilst donating the proceeds to the guerrilla organization. Recent trips to the west coast, however, piqued her interest in and the many shapes and forms this takes already now today in our daily lives. This then led to her idea of co authoring the book with ChatGPT. Rising AI tech demystified. And what I love most about her is her life motto. Life is not about fitting in, it's about standing out. And so, in her perfect red outfit, today, she's the right candidate for a podcast on a crowded Money 2020 floor.

Speaker C: Uh, Angelique, thank you for this, uh, introduction. Ah, corny. So let's dive in. Uh, Angelique, um, we just heard that you made a trip to the US to learn about AI and machine learning, which triggered the book. So tell us about your experience, please.

Speaker B: Yeah, I think the reason goes back really long ago. Um, when I was a little girl, I always played with Meccano, with technical toys. So I've been fascinated by hardware, but also software and techs. I wanted to build. Um, so last year I traveled to Seattle and the bay area with 25 Dutch entrepreneurs and investors, and we really wanted to learn about machine learning, data science and AI. Um, so we started in Seattle and all the Bay Area, and then there was one place that we wanted to visit, but we couldn't find the entrance. We were completely lost walking in San Francisco. And the great thing is we were on our way to OpenAI and by that time, ChatGPT didn't even launch yet. We couldn't find the entrance. So imagine 25 people walking around wondering, and we were like, how can we miss a 96,000 square foot building? And we're like, oh, this was going to be the highlight. Um, but then we walked in and rushed to the meeting room and all of a sudden we saw Sam sitting there and Lady Shell shouted us, no pictures. And we were like, oh, shoot, too late. Um, and we got a preview and all of us were like, wow, you know, AI is not this new thing, but giving it in a UX way and how easy it was to use. We were like, blown away. So I traveled back home after that trip. 40 different companies visited and felt I need to put this to the test. How can I do this? And beyond the simple prompt. So that is how I came up with the idea. Let's write a book with a AI co author instead of a human co author, which I didn't before.

Speaker C: Yeah, and why were you triggered in the first place to, to, uh, implement, investigate AI. Why did you take the trip?

Speaker B: Um, if you look at some of the highlights in the past history about tech innovation, first we had cloud, uh, then we had low code, no code. And AI has been like this almost underground trend which a lot of companies embrace. Machine learning, um, and then AI, you see so many more use cases starting to ignite. And for us it was like, okay, it's such a topic that everybody says something, they know about it, but what is it actually? And there are not a lot of people who are deep specialists. So we thought that let's go to the US and really talk to these people. Because we're not specialists as well.

Speaker A: No, I think exactly is that undercurrent is so important because what I see is that because it's now on literally the front page of even the dailies on a daily basis, um, a lot of people talk about AI, but we throw all of the aspects in one pool. So I'm really like, do you actually feel that people are aware of what AI is, or do we need way more education on the topic? Um, because I think a lot of things we spoke about are really what you say are machine learning or their optimization. But, you know, artificial intelligence has something to do with the neuro. The neuro element.

Speaker B: Yeah, good question. And, uh, I'm not the biggest AI specialist I know use cases. So, um, you see so many people saying that they know something about AI, but it all begins with data and data science. And a lot of companies are still struggling to get that in place. So I think that is the major core about it. And then, of course, you have machine learning. And within that circle, there's already so many knowns and unknowns, and then exactly what you say then comes in AI. And how do you, uh, define, uh, AI? Because you have generative AI, you know, creating new content. Um, and it's been around for, I think over six, seven, eight years. Um, all these models. But is education needed? Yeah, but where do you get it? Uh, I think that's the question. Um, what of the examples in the US if you've studied data science and are, are an expert in machine learning, you get a starting salary of $500,000. It's unbelievable. But that says something about how scarce this real deep knowledge still is. Yeah, so, yeah, a lot of education needed and experimentation. And experimentation. But yeah, even, you know, I speak to a lot of CEOs, uh, of financial services institutions, and they don't know it as well. They don't have an understanding on what um, it is, how to utilize it. And they do have people working on it, but a strategy or uh, uh, it implemented throughout organization.

Speaker C: Yeah.

Speaker A: And I think statements like uh, and they all make them for their own reasons, but statements like Goldman Sachs makes where he says AI will possibly replace 300 million jobs. I don't think it adds a lot of, of decent content to the debate because it really just scares people off.

Speaker B: Yeah, I think the biggest thing is that I call it the Xerox line. Um, in the past, everybody below a certain generation, uh, uh, content generation, it was consultants creating and copying uh, slides and it was just repetitive work. And I do think that AI will replace everything below the Xerox line. But what is then the difference and where will it have the impact is above that Xerox line. So where creativity comes in and where new content generation comes in and yeah, are executives focused on that? Do they understand that there is a Xerox line and everything below? Yes, jobs will be replaced. But I'm not saying disappear, I say replaced.

Speaker A: Yeah, exactly.

Speaker C: So we're going to talk about the society risk later on on in the podcast. Um, uh, first of all, I was um, triggered by your educational experience. Um, because you know, you co authored the book. Um, it was an educational experience for yourself, but hopefully also a learner for others. So take us on your journey, please. So how, how did you go about? What prompts did you use to create a true deepening dialogue?

Speaker B: Well, most importantly, we are now in June, June 2023. So I used the version before December. So that's a big thing. There were no threats, There was not a paid version. Um, GPT kept crashing like over and over again. So I was trying to build my.

Speaker A: It was almost like a real human being.

Speaker B: Almost the same.

Speaker A: Exactly.

Speaker B: I had trouble getting a bond with my new uh, co author because I had to rebuild that relationship every single time because, because it crashed. So I lost my history, which meant that I had to. It's like Groundhog Day of dating with AI over and over again. So that was very frustrating. Um, but the first bit was really, really quick. You know, within a couple of hours I had the whole book outline and that would normally take me months. Um, and the quick generation up to let's say 2,000 words, 3,000 very fast. But then deep diving, I got so much repetitive content. Um, again, crashes. And at one point I got hallucination and yeah, not me. Don't look at me. I didn't know I'm Dutch. But no,

Speaker A: exactly what were you smoking?

Speaker B: I literally asked, give me a list of, uh, five authors who wrote about Ikigai in insuretech. And I got a list with five male names. I'm like, oh, again, male names. So I said, give me female authors. And I got another list. I'm like, okay, now we're good. And I looked at the titles, and I'm like, shoot. One title is almost the same as the other one. So I asked ChatGPT as if it were a human. Hey, tell me about it. Did they write the book together? And ChatGPT came back, said, no, no, it was the dude. I'm like, okay, M. And I pushed for it. Are you sure? And then I got the answer. A double apology from GPT. I'm so sorry. I apologize. Blah, blah, blah. And I'm like, apologize? The thing is apologizing to me.

Speaker A: Yeah.

Speaker B: Oh, no. So I said, okay. Are you really sure? Yes. It's the guy.

Speaker C: It's not a dude. Then chatgpt.

Speaker B: No, no. But if you apologize to me,

Speaker A: you're saying it, Brian. You're saying it.

Speaker B: But in the end, it turned out that neither of the two actually used GPT or wrote the book about Ikigai. And I made this huge mistake, and I didn't realize it. I think so many of us still don't realize it. What was my mistake? Do you think you believed him? Yes.

Speaker A: You didn't check.

Speaker B: I used it as a search engine. Oh.

Speaker A: Yeah.

Speaker B: It is not a search engine. It's a large language model. So that meant at that moment, I'm like, oh, no, this hallucination thing is real. I had to check every single fact in that whole book.

Speaker A: Yeah.

Speaker B: Uh, everything. But I had to stick to my principle of 50% by ChatGPT and 50% by me, which meant I had to compromise on certain parts of the quality.

Speaker A: Yeah.

Speaker B: And that, as an author, doesn't feel good because I never did that with my previous books.

Speaker A: No.

Speaker B: But it was more a psychological process. And I got totally confused. I was like, literally in an argument with the thing.

Speaker A: Yeah. Yeah.

Speaker C: And did you also, um, use, uh, version, uh, four already? And, uh, is that probably way better? Right.

Speaker B: Of course.

Speaker C: Yeah. Okay.

Speaker B: Of course. I think version 4 is really another breakthrough, not only in terms of parameters, but the example I like the most is one of the, uh, co founders of Open, and he drew on a piece of paper, like a wireframe from a website. He took a picture and it was converted to software code.

Speaker A: Yeah.

Speaker B: And it actually worked. You could push the buttons. So that's not my business. What I do, it generates, but it really generates it. So I think, you know, my future idea would be, oh, if I have an idea, because I have a creative mind, I can just talk to it and it will generate whatever I want. The whole body, business idea, the proposition, the apps.

Speaker A: Yeah.

Speaker B: And that is the bit that I really, really get excited about.

Speaker A: Yeah, let, let's dive a little bit then, because I, I love the whole experimental thing. And you're obviously a smart woman. You knew what you were getting into. You're, uh, not dependent on it. Um, but a lot of people are not in that mindset. And they will, you know, become part of the whole AI experience. In fact, it will touch everybody's life. So let's talk a little bit about the risks. Uh, you talked about business risk. There is copyright society risk, there's ethical risk. But let's, um, uh, take on the first two ones first. Businesses in general, they thrive by invention and innovation. Where do you think that AI will help or hinder us?

Speaker B: Yeah, I think, um, if you look at a company, um, you see a lot of C level executives, they have AI as a agenda point on the agenda of the board meeting. And to me that is already where it goes wrong. Yeah, it should be one of the things, a topic that you do discuss throughout your agenda.

Speaker A: It's the undercurrent of your business.

Speaker B: It's the undercurrent. It's, it will be, it will impact every single process. Um, but in terms of risk, um, I think it's a huge opportunity for them because now the real creativity and the real uniqueness of thoughts, that will be the unique selling point. That is how your customers will experience value.

Speaker A: But will the uniqueness, um, any innovation come from AI or will it still come from people, but it will be better expressed or better researched or better, better applied?

Speaker B: Yeah, good question. I think both. I think it will increase the speed of delivery because content is basically democratized at this moment.

Speaker A: Yeah.

Speaker B: Uh, you can get up to that Xerox level super fast. So I think it will increase speed. Will that oppose a threat? Yes. For certain industries, like the consultancy business? Yes. Will it offer an opportunity? Again? Yes. Will it offer a, uh, uh, challenge? Yes. Because how do you train and teach, for example, young potentials?

Speaker A: Yeah.

Speaker B: In the past they would walk and work next to senior consultants and try to make, uh, the presentations, write the reports, and that is how they learned. But now they will have to learn in a completely different way. And are, is management aware of that? How can I train, guide and manage these young potentials. So that is a challenge for them.

Speaker A: Them. Yeah. And, and another one, um, uh, on the copyright, um, normally when you write a book, the copyright, it's yours, it's, it's all yours, you know, so that is a new way of thinking as well. And of course as an industry we're already looking at open source and that collaboration is much better than owning stuff. Um, but how do you look at this whole uh, copyright and risk element? Because I think as an industry we might, might have to really rethink our uh, business models. Because so far even if you look at, you know, I'm heavily involved with VC funding, it's always about intellectual property, right? It's about property, right, yeah.

Speaker B: In terms of risk, there are plenty of risks. Um, if I compare it to the art world, there was a big discussion here in the Netherlands that there was a Vermeer painting, uh, replaced by an AI generation generated painting. And it was almost like a popular revolt that people were so angry about it. But to me, well, you know, it is, uh, if I look at the artists, what I don't agree with is that all these models are trained based on their production, on their data, on their creativity, but they do not get paid. There's no remuneration to the original, uh, copywriters, to the original artists. And that is what we saw also when Napster came around and all these music streaming platforms. So I think we're in that phase that we have to figure out how do we uh, reward and remunerate the original content creator, the original uh, artists? Because that is what is needed to feed these models. Yeah, um, and have we sorted that out? Not yet. But do we have to sort it out? Yes. Could potentially blockchain help with that? Because that is where ownership is really embedded. Yeah, maybe.

Speaker A: I think it's a good idea.

Speaker C: So talking about risk, and of course we shouldn't talk about risk all the time with the ChatGPT, because it's an opportunity too. Right, but, but let's talk about society, risk. And um, I'm not talking about the elections or deep fakes or those kind of things. You call it a line and other people call it creative, uh, desert. Right, so you have, um, now let's call the average, uh, people, they're all on the same level, right? Because the old text is uh, generated from the new text, is generated from the old text. So there are a few people who can, you know, natively and from a human perspective create text. So don't you see a very Big gap between the people who are really creative and the rest of us.

Speaker B: Um, the question is, is that difference already here? Because if I would ask you the question, do you know how GPT is trained? What five sources are used? Do you know it? Do you know any sources? Yes.

Speaker C: Well I know they take the old text, right. So whatever is out there, it's used.

Speaker B: Yeah. So I'll give you two examples. One of the sources is common crawl. It scrapes the whole intel Internet. So if original content is already represented on the Internet, that is already verbalized and visual within the model. Um, another source is the English Wikipedia. Um, what a lot of people don't know is that only 20% of the contributors are women. That means that the data and data bias to my opinion is the biggest risk. So would you call that a straight societal risk? I would. I think the data bias is, is something that every single governmental body, that every single financial services company, but all companies should be aware of. And to me that will be priority number one because data is going to commoditize and in the end that will lead to that desert that you're talking. If we don't fix that underlying layer then yes, the desert will get bigger and bigger like we're doing now already to Earth. But true creativity will rise above. But not necessarily through the models.

Speaker A: No.

Speaker C: And who should fix it? By the way?

Speaker B: Uh, that's a very interesting. I uh, saw a presenter yesterday, um, and she also said there is not one government or overarching uh, uh, um, point I compare to stuff Star Trek. You know in Star Trek you have the Prime Directive that is agreed upon like how do you interact with new societies? How do you show and give them technology? And I think that is what is needed for AI. But who should do that? I think it's a collaboration across uh, government and public um, sector but also with companies and businesses.

Speaker A: Yeah, yeah. This moves already in ah, the, in the direction of the, the ethics and what I strongly feel about this and it's on many other topics as well. But again we're having a debate on something which has massive impact on the future with still few people and we're deciding on the future of literally everyone. How do we indeed as a society deal with this whole ethical debate? Because it also has the potential to make life so much, much better. Better for people who have less access.

Speaker B: Yes, um, it is a challenge because uh, the Western societies are very verbal about this. Like the 1000 entrepreneurs all signing like we need to hold this for six months. But how about the People who are less close to AI and the use and the benefits of those models, I think that they are a little bit overseen at this moment. So I do agree with that. That's a huge, huge risk. But there also, there is a big upside. So, um, is it ethics? Yes. I think a lot of these big companies do have agencies and teams for it. Um, but there's also an upside. I don't know. Have you heard of a Dutch company called Pera?

Speaker A: Yeah.

Speaker B: Yeah. So I think Pera is a great example. They focus on using AI in um, a flexible model and they use it for recruitment and for people development. But the companies using a solution like that in terms of diversity, it increases by 33% because it takes away complete bias, free hiring. And those solutions are already there. So I think from ethics perspective, I'm an optimist. So I look at the Haas Glass full, um, if that is already here, one third gain within diversity within a company, that's a massive impact. So, yeah, is it a challenge? And, um, should every company worry about it and focus on it 100%.

Speaker A: And then one final one on this ethics thing, what is the one thing, uh, that you feel we should not worry about with um, AI people to say, oh, it will take over everything, it will do everything. I for one think I cannot see how AI can ever be be empathetic apart from in its tone of voice. What is your optimistic view on what it can never replace?

Speaker B: Oh, yeah, I'm a Trekkie, so I think of Data, um, the character in Star Trek. He always wanted to be human and he was not. Um, but he was so real like and real life that would I be able to define what is human and feeling? I, I can make a proximity, but I would never dare to. I, um, think there is a bigger challenge, to be honest. I think even talking about rules and regulations, um, now, for example, antitrust regulation is price based. I believe it should be privacy based because my worry would be that, um, uh, data and data and privacy will be something for the rich because the less data you actually share with the world, probably the higher the prices you have to pay. And to me that is an ethical concern. But again, I'm an optimist. You know, I believe that technology can be used for good.

Speaker A: Yeah, well that's, that's good news. Actually. This morning as I was driving to Monday 2020, I heard that, uh, one of the major, um, international awards has been given to a Dutch female lawyer who for the last 30 years has specialized in, in digitization and law. And she is, um, very active in the domain of, um, data privacy and um, the whole domain of, you know, how do we as lawmakers and as society deal with the rights of every single individual in that context? So I thought that was a really uplifting, um, sort of piece of news this morning on the radio.

Speaker C: Yeah, yeah, I think that's uh, the next, uh, the next guess. But because again, Connie, we managed again to get a topic that we can cut up in 10 different podcasts. But all good things must come to an end and in particular also this one at, uh, Money 2020. So before we go out, uh, Angelique, maybe the 10 million dollar question, and maybe in the light of Sam Altman and Geoffrey Hinton, is technology going too fast or are we humans going too slow?

Speaker B: Good one. Well, uh, in terms, we always go too slow. That's my perspective. Um, I compare where we are now to when humans got to experiment with electricity and how do we implement electricity within every single aspect of our life. And I compare that to AI. How do we implement it? It has such an impact that to me, you know, comparing it to a mobile phone is not even big enough. I would compare it to electricity.

Speaker A: Well, for people wanting to lecture themselves and others on anything FinTech, Innovation or AI, I suggest you follow Angelique closely on Twitter, on LinkedIn and @Brian.

Speaker C: Angelique, for real and curious which music and I know know that Angelique likes, uh, heavy metal. Check out www.fintechcapuccino.com AngeliqueShouten. Angelique, thank you for joining us here at Money 21st booth for a moneypot Fintech cappuccino.

Speaker A: And thank you all for listening to this very special Monipod cappuccino. If you don't want to miss another cup, subscribe to our podcast via Spotify, itunes or wherever you'd like to listen to to it. And please give us a like or a review so many more people can find us.

Speaker C: And please join us again on Saturday morning at 9. We'll have the coffee rated just the way you like it.

Speaker A: And all of you on air, online or at the show, enjoy Money20 20. So much to learn and explore and a curious mind is a joy forever.

Speaker C: Thank you Angelique.

Speaker B: Thank.

Speaker C: You.