OpenNashPodcasts

Analyse Asia with Bernard Leong

Dow Jones & AI in Risk and Compliance with Joel Lange

31 Oct 2023 · 38m · Joel Lange, Bernard Leong · 22 quoted facts

▶ Play episodeOriginal source ↗

In brief

Joel Lange of Dow Jones argues that fighting money laundering and financial crime depends on pairing structured risk databases with trusted journalism, and that AI now lets banks search both at high volume. Dow Jones's new screening tool, ASAM, combines its risk data with the Factiva news archive. Lange says the aim is to catch more real risk while wasting less time on false alarms.

The central tension is between being effective (missing nothing) and being efficient (not drowning in false alerts). Lange says regulators long rewarded wide nets, which created floods of false positives that hid real risks. AI promises sharper answers, but he warns that generative AI must respect journalistic integrity, copyright and data privacy, so firms should not break rules to enforce rules.

Key insights

  1. Dow Jones's risk business began with a dictator's loot

    Lange says it grew out of Nigerian ruler Sani Abacha's looting in the 1990s, which led major Swiss banks to ask for a database of politically exposed people. Sanctions lists and criminal profiles were added later.

  2. Wide matching nets can make compliance less effective

    Lange says regulators once seemed to want many false positives as a safety blanket. That volume created inefficiency, which can cause teams to miss real risks.

  3. AI lets banks ask questions and get answers

    Lange contrasts older screening, which returned many possible matches, with AI that uses extra details such as dates of birth and countries to narrow results. That makes high-volume checking of news more practical.

  4. Screening must cover customers, ongoing monitoring and payments

    A customer can be clean at onboarding and later appear in sanctions or adverse media. So Dow Jones data is used when onboarding, for ongoing monitoring, and in filters that check payment beneficiaries.

  5. Scraping journalism to fight crime creates new risk

    Lange says firms should not scrape news from the internet to power compliance tools, since that can violate copyright and intellectual property. He cites growing respect for journalism in laws in Australia and Canada.

Questions

What does Dow Jones actually do in risk and compliance?

Dow Jones sells databases of politically exposed people, sanctioned individuals and entities, and profiles of people charged with financial crimes, plus a news archive called Factiva. Banks and companies use them to spot high-risk customers and counterparties.

What is adverse media and why does it matter to a bank?

Adverse media means negative news coverage, such as a customer being linked to fraud or money laundering. Lange says that if a bank is found to be working with such a person, the reputational damage can be significant.

What is ASAM and how is it different from older screening tools?

ASAM is Dow Jones's new Risk Center advanced screening and monitoring tool, which brings its structured risk data and the Factiva news archive together for high-volume checks. Lange says older rules-based engines struggled with unstructured data like news articles.

Why do false positives cause problems in anti-money laundering checks?

A false positive is an alert that turns out not to be a real match. Lange says too many of them create inefficiency that can lead staff to miss genuine risks.

Can companies use generative AI for compliance work safely?

Lange sees big efficiency gains, such as mimicking how a researcher follows leads and summarising findings. But he says firms must protect journalistic integrity, respect data privacy, and guard against misinformation or hallucination.

Facts

Stated by the speakers. Each quote matches the transcript, which shows what was said, not that it is true.

Joel Lange said that after Sani Abacha died in 1999, his wife was found at an airport with suitcases full of cash.

“when he died in 1999, his wife was found at the airport with suitcases full of cash”

0:09:18 · Joel Lange · Sani Abacha · 1999

Joel Lange said that in the 1990s Sani Abacha and his relatives and associates extracted cash from the Nigerian economy and placed it in European banks.

“Abacha and his relatives and close associates had been extracting cash from the Nigerian economy and effectively placing it in European banks”

0:09:18 · Joel Lange · Money laundering · 1990s

Joel Lange said major banks, particularly Swiss banks, asked for databases to identify politically exposed people, which led to Dow Jones's risk and compliance business.

“we want to be able to identify political exposure, to effectively identify politicians and their relatives and their close associates.”

0:10:07 · Joel Lange · Business origins

Dow Jones added sanctions data, sanctioned entities and profiles on criminality to its original political exposure databases, according to Joel Lange.

“we've continued to add to those databases with sanctions and sanctioned individuals, sanctioned entities, as well as profiles on criminality”

0:10:07 · Joel Lange · Risk data

Joel Lange named the Monetary Authority of Singapore, the Hong Kong Monetary Authority and the US Office of Foreign Assets Control as sanctions authorities that firms must follow.

“sanctions regulations here in Singapore with MAS, in Hong Kong with HKMA, the US government with the Office of Foreign Asset Control”

0:11:05 · Joel Lange · Sanctions regulators

Joel Lange said banks have been fined billions of dollars over the years for sanctions failures.

“We've seen fines over the years in the billions of dollars for various banks.”

0:11:41 · Joel Lange · Regulatory fines

Joel Lange said Dow Jones's risk and compliance business is based in London with its global research center in Barcelona because of heavy cross-border financial traffic.

“why our business is based in London and we have our global research center in Barcelona is because there's so much cross-border traffic”

0:15:43 · Joel Lange · Company locations

Dow Jones data is built into payment filtering solutions so that banks can check outgoing payment beneficiaries against sanctions, according to Joel Lange.

“we slot our data into a lot of the payment filtering solutions, so that the banks can check any payments going out the door”

0:15:13 · Joel Lange · Payment screening

Joel Lange described Factiva as a global aggregation of news that includes The Wall Street Journal, Reuters and other publications.

“Factiva.com, which is a global aggregation of news. So it includes the Wall Street Journal, includes Reuters, includes various publications”

0:18:45 · Joel Lange · Factiva

Dow Jones chose UK company Ripjar's technology for its advanced screening and monitoring tool to distil the Factiva news corpus and extract names and metadata.

“We chose a technology called Ripjar out of the UK, which really does some fantastic things in relation to taking the corpus of Factiva, distilling it down”

0:19:27 · Joel Lange · ASAM technology

Joel Lange recalled that in 2008 the New York banking regulator tested banks by entering names with transposed characters.

“back in 2008, the New York regulator going into New York banks and actually entering in transposed characters”

0:22:20 · Joel Lange · Regulatory testing · 2008

Joel Lange said some organizations set arbitrary fuzzy-matching thresholds, such as 80 percent, to decide whether a name matches a sanctions entry.

“organizations may kind of set like an arbitrary percentage level to say anything below 80 % won't match, anything above 80 % will match”

0:23:20 · Joel Lange · Name matching

Dow Jones translated the US OFAC sanctions list into Chinese characters and then Chinese commercial code for customers in Chinese-speaking regions so payments could be matched number to number.

“they actually wanted us to translate the list from English characters into Chinese characters. And then in addition to that, take those Chinese characters and convert them into Chinese commercial code.”

0:25:22 · Joel Lange · Sanctions translation

Dow Jones built a specialised organized crime data set for Japan, according to Joel Lange.

“in Japan, we focused on organized crime. There's been a specialized data set that we've worked on there.”

0:25:58 · Joel Lange · Japan data

Joel Lange said Dow Jones regards the Singapore regulator as one of the leading regulators worldwide for giving prescriptive guidance.

“We actually look at the regulator in Singapore as one of the leading lights around the world”

0:26:28 · Joel Lange · Singapore regulation

Dow Jones built a sanctions control and ownership database of entities owned or controlled by sanctioned individuals, importantly after the invasion of Crimea, according to Joel Lange.

“We actually, almost 10 years ago, built a data set called sanctions control ownership, which essentially was a database of entities that are owned or controlled by sanctioned individuals.”

0:27:49 · Joel Lange · Sanctions data

Joel Lange cited recent articles reporting that Hamas used cryptocurrency to fund its programs.

“recent articles talking about Hamas using crypto in relation to the funding of their programs”

0:28:28 · Joel Lange · Terror financing

Joel Lange said Dow Jones's parent company is News Corp, which collectively owns major journalistic assets.

“our parent company, News Corp. Collectively, we own some of the best journalistic assets in the world.”

0:29:58 · Joel Lange · Corporate structure

Joel Lange said Factiva pays royalties to newspapers around the world that contribute their publications to it.

“we effectively pay royalties to newspapers around the world that contribute their publications to Factiva”

0:30:32 · Joel Lange · Factiva royalties

Joel Lange began his career in financial crime data at Acuity, a company divested from Thomson Financial, before moving to Dow Jones.

“Started working for a company by the name of Acuity, which had been divested from Thompson Financial.”

0:03:50 · Joel Lange · Career

Joel Lange moved to Cologne, Germany in 1999 on a US government program.

“And I moved in 99 to Cologne on a program through the US government.”

0:03:50 · Joel Lange · Career · 1999

Joel Lange described Wall Street Journal reporter Evan Gershkovich as a wrongfully detained colleague and said people were working to secure his release.

“having a colleague, a friend of many of our colleagues here at the Wall Street Journal and Dow Jones wrongfully detained.”

0:02:39 · Joel Lange · Evan Gershkovich

Chapters

  1. 0:03:50From German immersion school to financial crime data

    Lange explains how a German-language upbringing led to a career in anti-money laundering data. It frames his long experience in the field.

    “I always wanted to live in Germany. And I moved in 99 to Cologne”
  2. 0:06:45Dow Jones's mission: unique, ubiquitous, united

    Lange outlines the company's pillars and how the risk and research business identifies both risks and opportunities.

    “we have the core pillars of being unique, ubiquitous, and united”
  3. 0:10:07How an African dictator's loot created a data business

    The Abacha story explains why banks asked for databases of politically exposed people, the origin of the risk business.

    “the major banks, particularly the Swiss banks, got together to say”
  4. 0:11:41The cost of failing to manage risk

    Lange separates regulatory risk (fines) from reputational risk (bad press) and points to sanctions regimes in Singapore, Hong Kong and the US.

    “We've seen fines over the years in the billions of dollars for various banks”
  5. 0:13:04Balancing effectiveness and efficiency across workflows

    Lange describes how data is used at onboarding, in ongoing monitoring, and in payment filtering, and why cross-border flows raise risk.

    “what we really focus in on is balancing effectiveness and efficiency”
  6. 0:19:27ASAM: merging risk data with Factiva news

    The new tool pairs structured risk data with the Factiva news archive using AI technology from Ripjar to screen high volumes of names.

    “We chose a technology called Ripjar out of the UK”
  7. 0:22:51From exact matching to fuzzy logic to AI

    Lange traces how regulator expectations and matching methods evolved, and why arbitrary thresholds and false positives became a problem.

    “it was very much exact matching”
  8. 0:25:58Asia-Pacific specifics: Chinese codes and Japanese data

    Regional needs such as translating sanctions lists into Chinese commercial codes show why local language and regulatory knowledge matters.

    “they actually wanted to match number to number rather than name to name”
  9. 0:27:49Sanctions on Russia and global fragmentation

    Lange explains how a database of entities owned or controlled by sanctioned people helped firms assess their Russia exposure.

    “a database of entities that are owned or controlled by sanctioned individuals”
  10. 0:32:59Generative AI, journalism and data privacy

    Lange lays out Dow Jones's three roles and advises firms not to scrape journalism to fight financial crime.

    “don't commit a financial crime to combat financial crime”

Transcript

0:00:01
Joel Lange

Do you manage your own IT for distributed teams in Asia? And you know how painful it is. S-A -V-L helps your in-house team by taking cumbersome tasks off their hands and giving them the tools to manage IT effectively. Get help across eight countries in Asia Pacific, from on - and off-boarding, procuring devices, to real-time IT support and device management. With our state-of-the-art platform, gain full control of all your IT infrastructure in one place.

0:00:30

Our team of IT support pros are keen to help you grow. So check out esevel.com and get a demo today. Use our referral code ASIA for three months free. Terms and conditions apply. And so what we do at Dow Jones very specifically is look at what are the regulators demanding? What do they want you to do? what are the specific data points that you require and which workflows do you want to use them in? And there are very different workflows which within, for example, a financial institution, you have the onboarding of your new customers.

0:01:08

And in those situations, you may have a little bit more time. Obviously, the service level in terms of opening a bank account is faster and faster, but you have a little bit more time in that regard. You want to look at potentially more data. Then once you are a customer of a bank, you're going to be amongst millions of customers. How do banks then adapt to check additional data points? And today, the customer may be clean and have no adverse media, may not be sanctions.

0:01:36
Bernard Leong

That could change, right? Welcome to Analyse Asia, the premier podcast dedicated to dissecting the pulse of business technology and media in Asia. I'm Bernard Leong, and while we are coming to grips with AI, with the rise of chat GPT, mid-journey and stable diffusion, it is important that we should put things in context about artificial intelligence. One key area is the use of artificial intelligence in risk management and combating financial crimes. To help me understand how this is done, I have Joel Lange, Executive Vice President and General Manager of Risk and Research from Dow Jones.

0:02:17

Joelle, welcome to the show. Thank you, Bernard. It's great to be with you today. Today is a very interesting date because the Wall Street Journal is part of Dow Jones. And I think you have a very well-known journalist. Actually, I know it because one of my friends who I've interviewed on this show, Nouni Purnell, has always been constantly putting up pictures. Today is the 32nd birthday of Evan Gershkovich, right?

0:02:39
Joel Lange

It is, yeah. And it continues to be a sad and challenging situation. having a colleague, a friend of many of our colleagues here at the Wall Street Journal and Dow Jones wrongfully detained. And there are people working around the clock to hopefully have him released sometime soon. And it is a real tragic situation in that regard. But we're all hopeful for Evan's release and continue to work hard towards that.

0:03:09
Bernard Leong

And we wish you all the best in trying to get him released from Russia. So we want to know your origin story. How did you start your career and eventually land up in your present position with Dow Jones?

0:03:20
Joel Lange

Well, I guess we all have our stories. And I think for me, I grew up in the Midwest and in the US. I happen to be from a family that had some German origins and grew up attending a German school and learning German from the age of four, an immersion program where all my lessons were in German. And that in a way kind of brought me to where I am and that I eventually went when I kind of tested through German and high school went on to university.

0:03:50

I always wanted to live in Germany. And I moved in 99 to Cologne on a program through the US government. Actually met my wife in Germany, who's from the UK and ended up going over to the UK where I did my master's and in a way kind of fell into the world of anti-money laundering and financial crime. Started working for a company by the name of Acuity, which had been divested from Thompson Financial. It was the former Thompson Financial Publishing.

0:04:19

Started working in sales there and absolutely loved the industry of payments and anti-financial crime data. And over the years have really evolved into working in sales roles, being passionate about product, working then in product management, moving into kind of executive roles as I moved on from Acuity to Dow Jones, and have continued now for over, really over 18 years in the world

0:04:48
Bernard Leong

of payments and financial crime. And one important question I usually ask, because there's a lot of younger audience who's actually listening to this podcast, can you share the most important lessons in your career journey? Well, yeah, I've been really fortunate to have a lot of great mentors

0:05:04
Joel Lange

as well as experiences through that journey. And I remember early on in my career, really going through stressful situations and times where you're pushing towards achieving goals, having interactions with customers and colleagues. And you get an email late at night that is kind of emotionally charged. And I think I had a mentor early on that just always told me, look, it's always important to sleep on it. You get an email late at night that's emotionally charged.

0:05:33

is make sure that you remain calm and sleep on it. And I think that's always been a great bit of advice for me. I think in addition to that, I've had mentors throughout the years have talked to me about hiring and looking for good people and really what makes people successful in this industry. And I think that I heard it best kind of characterized that those that do well have an engine and they also have intellectual bandwidth.

0:06:00

And the way that I've interpreted that over the years is obviously the engine is clear, right? You need to work hard. You need to be motivated. The intellectual bandwidth piece, I think is just a combination of curiosity and creativity. And I think that, you know, the people that I see in this space are doing well that, you know, again, have that baseline engine. They're very motivated, but have, you know, that curiosity and creativity, I think is definitely something that I look for definitely

0:06:28
Bernard Leong

when I'm hiring as well. And it comes to the main subject of the day. I want to talk about Dow Jones and also using artificial intelligence or AI in risk management and combating financial crimes. So maybe to start and help my audience on the subject, can you share the global vision and mission of Dow Jones and what it does?

0:06:45
Joel Lange

Well, Dow Jones, you know, being an information provider and also the owner of various journalistic entities, we're very much focused on providing trusted journalism to help organizations make decisions. And in relation to that, we have the core pillars of being unique, ubiquitous, and united. And so just kind of double-clicking on that, unique that we have very unique assets in terms of the trusted journalistic pieces that we put out, as well as the content and databases that form our offerings.

0:07:20

Ubiquitous, and we're a global company, we're pervasive throughout the markets in relation to our products and services. And United, that's something that has really been a key part of our management team's focus is that we have offerings that are direct to the customer. We have offerings that are to organizations, but ultimately it's business professionals that are making these decisions and we look to serve them. I think more specifically in terms of my role, looking after the risk and compliance business, as well as our research arm.

0:07:55

We call it the risk and research business. But in many ways, what we do is very much focused on the identification of risk, various risks, but also finding opportunities. So when I look at the various use cases and the various customer challenges that we're looking to meet, we're helping our customers to effectively identify various risks, be that money laundering risk, anti-corruption risk, capital markets risk. On the other side of things with Factiva and with our information and the wider media, we're able to help identify opportunities.

0:08:35

And speaking to a colleague today, obviously being here in Singapore about the Chinese character for crisis also includes opportunity, right? And so they talk about Wei Z and that character, I think that kind of encapsulates what we do in relation to the

0:08:52
Bernard Leong

identification of risk and opportunity. I think for myself, who have actually been the head of retail for Singapore Post, where I actually have to handle remittances, so I understand what risk management, anti-money laundering, and also financial crimes. But for the greater general audience out there, I think it may be good. Can you define what constitutes risk management, anti-money laundering, and financial crimes to level set my audience in understanding this complex subject.

0:09:18
Joel Lange

Well, sure. And happy to do that. And the origin story around the risk and compliance business at Dow Jones goes back 20 years to a time over 20 years ago where you had essentially dictators in various countries that were some famous stories like, for example, Sani Abacha, when he died in 1999, his wife was found at the airport with suitcases full of cash. And it turned out that in the 90s, Abacha and his relatives and close associates had been extracting cash from the Nigerian economy and effectively placing it in European banks, placing it and then layering it and using these various money laundering techniques.

0:10:07

And so the origin story of Dow Jones Risk Compliance was that the major banks, particularly the Swiss banks, got together to say, we want to be able to identify political exposure, to effectively identify politicians and their relatives and their close associates. And, you know, we were approached as well as a few others to build databases on political exposure. And over time, we've continued to add to those databases with sanctions and sanctioned individuals, sanctioned entities, as well as profiles on criminality to build up a database that actually help organizations to identify high risk people and companies within their organization.

0:10:52
Bernard Leong

Can you describe the financial and reputational impact on organizations if they do not manage their risk well or fail to capture some form of financial crime that may be even happening within the organization that they are not aware about?

0:11:05
Joel Lange

Well, and that can happen in various different ways. And, you know, I think we often talk about the different types of risk that we're helping our clients to mitigate. And so there's obviously the regulatory risk, that risk of actually being fined. And so that is related very directly to sanctions. No sanctions are absolute, particularly for, you know, when you're working in a country that where you're you have to adhere to their sanctions regulations here in Singapore with MAS, in Hong Kong with HKMA, the US government with the Office of Foreign Asset Control.

0:11:41

And so it's those jurisdictions you work in, as well as the currencies you trade in. It's really important to be mindful of ensuring that you're not working with anyone that's on those sanctions lists. And obviously, the consequence of that can be really significant. We've seen fines over the years in the billions of dollars for various banks. The other side of it is then a reputational side. So obviously, a lot of the data that we work on and a lot of our modern solutions are focused on adverse media.

0:12:13

And so if it turns out that you're working with a money launderer, a fraudster, and that comes out in the press, that has significant reputational damage. Obviously, with various money laundering scandals, obviously, there's been a few here recently. You really don't want to be that organization that's in the news to have known that you're working with a potential fraudster or money launderer.

0:12:37
Bernard Leong

You're referring to the Chinese Fujian gang, which was actually apprehended recently for money laundering in Singapore, which is one of the major news here. So one thing I'm curious, what is the role and perspective of the Dow Jones Risk and Compliance Division, particularly in the realm of, say, managing risk and combating financial crimes? And what kind of challenges typically you've seen organization, businesses or even governments face specifically in this area?

0:13:04
Joel Lange

Well, it's a huge challenge for organizations. And what we really focus in on is balancing effectiveness and efficiency. And so those are kind of the buzzwords in our space, of course, but I think really drilling into them is important. So what do people mean by being effective and efficient? And I think what is really important to note is that often compliance teams are cost centers and obviously don't generate revenue. They are obviously trying to protect the bank or protect the corporation from a fine or a reputation.

0:13:40

but sometimes it's very intangible in that regard. But there's so much pressure to be effective. So really, are you catching the data points that you need to catch? No one wants to miss anything. And so it's really important to do that. However, you also need to be efficient. And oftentimes in the quest to be effective and actually look at all of these data points, you can create huge amounts of inefficiencies, which then potentially lead to you missing something as well.

0:14:07

And so it's a real balance. And so what we do at Dow Jones very specifically is look at what are the regulators demanding? What do they want you to do? What are the specific data points that you require? And which workflows do you want to use them in? And, you know, there are very different workflows, which within, for example, a financial institution, you have the onboarding of your new customers. And in those situations, you may have a little bit more time.

0:14:35

Obviously, there's the service level in terms of opening a bank account is faster and faster. But you have a little bit more time in that regard. You want to look at potentially more data. then once you are a customer of a bank you're going to be amongst millions of customers how how do banks can then adapt to check additional data points and you know today the the customer may be clean and have no adverse media may not be sanctions that could change right and then lastly is the payments that your your customer sends so you know your customer but what about the payments that you're sending out, you need to check against the beneficiaries.

0:15:13

And so oftentimes, our data is used in all those different workflows. So new customer, ongoing monitoring, and then also, we slot our data into a lot of the payment filtering solutions, so that the banks can check any payments going out the door to see that those beneficiaries are

0:15:30
Bernard Leong

not sanctioned. And it's also becoming very complex, because it is now global as well. Sometimes you may just slip into it without realizing it. But based on this, your solutions actually flag it up so that they're aware of that. Absolutely.

0:15:43
Joel Lange

And it's really the cross-border nature that's the risk. And going back to the origin story of Dajun's risk compliance and why our business is based in London and we have our global research center in Barcelona is because there's so much cross-border traffic. You have politicians in Italy going to open a bank account in the UK and vice versa. And we see similar paradigm in Asia, right? We see a lot of money coming into Singapore from around the world.

0:16:09

And there's a real focus in Singapore about checking the source of wealth, checking who these individuals are, do they have some political exposure? And so it's really important to look when you're doing that analysis of, you know, the country of origin, but also the country of citizenship. And those may be oftentimes, especially with high net worth individuals, very different.

0:16:32
Bernard Leong

So what's the one thing you know about risk management and combating financial crime that

0:16:36
Joel Lange

not very few do? Well, I think we have a really unique status at Dow Jones. I think it's really very, very different to anyone else that works in this space in that we have the Wall Street Journal, and then we also have this fantastic risk business. And so the last time I was in Singapore, I was here with our CEO, and we were off seeing the regulator talking about it was just after the guilt crisis in the UK, and we're talking about inflation, and we're talking about interest rates.

0:17:03

But in the same conversation, we're also talking about the guidance in relation to name screening here in the region. And that's a really unique perspective. You don't really often have that. I guess similarly, I had the pleasure of being in Davos last year with the Wall Street Journal, got to see Satya Nadella talking about ChatGPT. And then we're working on solutions to help our customers automate due diligence reports and use generative AI in relation to capturing risk.

0:17:31

And so we really sit at that intersection. So I don't know if it's necessarily something we know that others don't know, but we have this unique opportunity to really be right at the forefront

0:17:41
Bernard Leong

of inflection points on technology as well as regulation. I also know that recently Dow Jones has this newly announced risk center advanced screening and monitoring, which is actually based on the best-in-class risk data that you actually already have and also adding cutting-edge machine learning algorithms. Can you talk about the solution itself?

0:18:02
Joel Lange

I'm glad you brought that up. It's also very much at the intersection of my new role. I guess very recently, I was elevated to, in addition to looking after our risk and compliance business, to also look after Factiva. And as mentioned, we call that risk and research. And the new advanced screening and monitoring tool in many ways encapsulate those two solutions. So for years, customers have used the baseline structured risk and compliance data to identify politicians, sanctioned individuals, structured negative news, where we've actually built a profile on an individual that has been arrested, convicted, or charged of a financial crime.

0:18:45

In addition to that, they've also often bought a seat of Factiva.com, which is a global aggregation of news. So it includes the Wall Street Journal, includes Reuters, includes various publications, trusted publications, credible publications from around the world. And they've bought that additionally so that they could cast a wide net on negative media and media on the counterparty they're looking into. And what we've done with this new advanced screening and monitoring solution is really brought those worlds together and really leveraged a modern technology to perform high volume checks against both of these databases.

0:19:27

So 10 years ago, if you wanted to load millions of names or check millions of names against Factiva, it was quite difficult because although Factiva is highly structured, there's distinct tagging on all the articles, you were still using kind of old-fashioned Boolean searches to really hone in on the content. whereas we kind of scoured the market to look at all the new AI solutions. We chose a technology called Ripjar out of the UK, which really does some fantastic things in relation to taking the corpus of Factiva, distilling it down, but then also taking the ability for them with their entity resolution capability to extract names, extract metadata, which is really useful in the matching process.

0:20:11

And then also the AI is able to drill in and really understand the articles that are most relevant to that grouping and then surface that so that customers can really efficiently do their negative news screening.

0:20:25
Bernard Leong

So what makes it stands out, say, from other risk management solutions that exist in the market then?

0:20:32
Joel Lange

Well, it's really specifically focused on this, what we kind of call a new paradigm of screening and monitoring. So, you know, for years, there's been very much kind of rules-based engines that are focused on structured data And they have definitely got better and better over the years, but it's very much predicated on having a structured data set. And oftentimes unstructured data was more difficult to bring into that automated environment. And so where we're really setting ourselves apart is that with organizations that are now ready to, in a way, kind of identify the art of the possible and bring together these two content sets in an efficient manner, particularly for the negative news piece.

0:21:17

I think in future, there'll be even more data points that come in, looking at private company data, looking at other additional data nodes that could actually be consolidated in the profiles that are identified. But this step change of solving this problem of baseline negative news and expansive negative news is, we feel, truly unique in the market.

0:21:40
Bernard Leong

I think this is very interesting because unstructured data actually provides a lot of insights. And I think building that knowledge-based graph to actually try to put the structure and unstructured data together so it can actually get more insights specifically. So I think you have given a very good understanding of the AI and machine learning capabilities that how to refine that process. Can you talk about how the solution tackles, say, false positives, given that machine learning tends to provide solutions that are probabilistic.

0:22:09

And of course, there'll always be trade-offs in specificity and sensitivity of the data. I think this is something that you probably have put a lot of consideration and put into the product.

0:22:20
Joel Lange

Absolutely, absolutely. And I think going back to that, we very much kind of talk about the last 20 years and how solutions have evolved and with rules-based engines and also to an extent, the regulatory expectations. And I think that that is a really interesting piece to kind of drill down into. And so, you know, years ago, I can remember back in 2008, the New York regulator going into New York banks and actually entering in transposed characters, right?

0:22:51

Almost trying to catch out the bank to say, what if this happened? How would you deal with it? And so there was actually a movement. It's almost come full circle in a way is that in the early days of doing these types of checks, it was very much exact matching. But then the regulators were keen to ensure that organizations were using more variants. And so they would look at expanding algorithmically into a more Levenstein kind of fuzzy logic approach, looking at thresholds.

0:23:20

But that also, I think, had its pitfalls in that organizations may kind of set like an arbitrary percentage level to say anything below 80 % won't match, anything above 80 % will match without really looking into the context of, you know, what are the data points that could potentially make that match more sound? And I think, but, you know, the piece around really understanding variance, understanding the variance expectation of the regulator was always a challenge. It was almost like the regulator was saying, we want you to have lots of false positives so that it's almost like a warm blanket.

0:24:00

The more false positives, clearly they're casting a wide net. You won't miss anything. But again, as we talked about earlier, that inefficiency can actually lead to a lack of effectiveness. I think what we're seeing with AI, I guess the way I would characterize it is previously in those kind of first 10, 15 years of screening and monitoring, you would ask the system a question, you get lots of answers. I think the art of the possible with AI is that you're now able to really ask the question and get a response.

0:24:31

And the precision of that response is quite high if you're able to leverage a lot of the metadata that you have on the individual, right? And also, in many ways, certain attributes go beyond the standard metadata of name, date of birth, country, really looking at other pieces of information, if you're able to use them, that can help to disambiguate and really hone in on the answer to the question.

0:24:55
Bernard Leong

And I think with, of course, the generative pre-trained transformers, that will actually take another step up from trying to make sense of that data from what I'm alluding to what you just said. So specifically in the Asia-Pacific region, the regulatory landscape is pretty unique and challenging. One question I probably have is how does the diversity of the region, both economically and politically, actually impact the complex anti-money laundering processes?

0:25:22
Joel Lange

Well, I think there's just always a lot of specificity to the different markets. I think I go back a good few years. In my early days at Dow Jones, we identified that often for our customers in China as well in Chinese-speaking regions, that when they were looking at the U.S. government's Office of Foreign Asset Control sanctions list, they actually wanted us to translate the list from English characters into Chinese characters. And then in addition to that, take those Chinese characters and convert them into Chinese commercial code.

0:25:58

And that was because within payment environments, they actually wanted to match number to number rather than name to name. And we've seen that evolve over time. And that solution is now 10 years old. The precision in relation to the translation, the matching on the codes, I think that's been a key factor. We also see, you know, certain pieces like in Japan, we focused on organized crime. There's been a specialized data set that we've worked on there.

0:26:28

We've also worked on very specific politically exposed persons data in Japan. And then there's just, you know, the specificity of the regulation is keen. We actually look at the regulator in Singapore as one of the leading lights around the world. that we tend to see the regulator actually leaning into giving a lot of prescriptive guidance to organizations to what they should do. Often in some jurisdictions, it can be quite vague. So yeah, I think it's very much a mixture.

0:26:58

And as you say, the region is so diverse that it really takes having people on the ground with the linguistic skills to really understand not just the context, but also the matching paradigm. We just talked about the algorithmic approach. You bring in Chinese characters, kanji, katakana, hiragana. It's a complex area which requires

0:27:22
Bernard Leong

sophisticated solutions. Currently, with so much geopolitical events happening, I think about the recent Russian invasion of Ukraine or even the recent Hamas-Israel conflict. What is the mental model to strike the balance between, say, economic recovery and anti-corruption efforts, which actually affect the integrity of the financial sector in the countries, which sometimes can be tricky because of authoritarian governments in the region. So what's the mental model to

0:27:49
Joel Lange

actually look at this? That's a really, really good question. And I think there's a couple of different facets to it. When you look at the Russia-Ukraine crisis, really, and the data paradigm that came out of that. We actually, almost 10 years ago, built a data set called sanctions control ownership, which essentially was a database of entities that are owned or controlled by sanctioned individuals. And that was really important after the initial invasion of Crimea. And then most recently, with the most recent invasion, we saw a number of Russian oligarchs that were sanctioned.

0:28:28

They obviously owned a tremendous amount of assets around the world. And there was a big push post the invasion for organizations to identify their Russia exposure, whether they could do business, not do business, if they, you know, it went all the way through to property and yachts, and whether those those could be maintained, there was a whole license paradigm in relation to it. And, you know, I think equally that with geopolitical events, with terrorist financing, recent articles talking about Hamas using crypto in relation to the funding of their programs, the scrutiny that organizations have in relation to understanding their exposure and how they fall fall of what is effectively a global fragmentation that's happening right now across not just geopolitical

0:29:18

sanctions, but also in terms of supply chain and really identifying are your counterparties, are your suppliers, are even the customers, the importers of your goods, are they sanctioned or are they potentially, is there a more nuanced kind of trade sanction in relation to them, which is definitely a trend we see on the market.

0:29:39
Bernard Leong

So I think one question that always been cropped up so far in AI is of course, generative AI and its propensity to both augment, I think it should be fair. and also create misinformation in the form of context of, say, financial crime and risk management. What are your perspectives on that?

0:29:58
Joel Lange

Well, we have a number of perspectives. I think it is a challenge for us because we kind of sit in three buckets, I would say. So the first bucket, we are a publisher, right? And not just Dow Jones, but our parent company, News Corp. Collectively, we own some of the best journalistic assets in the world. And we are very keen to ensure that journalistic integrity is protected and how that data is essentially used within large language models.

0:30:32

And then the second bucket is with Factiva, we are effectively an arbiter of, we are kind of the compliance officer for the publishing industry because we effectively pay royalties to newspapers around the world that contribute their publications to Factiva. And then we use those Factiva. Factiva is used for multiple use cases. And there is a real demand from customers to use Factiva in a lot of these new generative AI use cases. And it's important for us to really take that seriously and communicate with our publishing community around how that content's being used.

0:31:08

But I guess the third is as a compliance technology, effectively a reg tech ourselves, we want to harness this technology to help solve customer problems. And it's so exciting in terms of the ability to bring efficiencies for those that are performing due diligence, particular on counterparties, the ability to aggregate information, the ability in a way to kind of mimic what a researcher does. So part of our business is that we do due diligence reports, for example.

0:31:39

And when a researcher does analysis on an individual or a company, they're effectively kind of, you know, looking into things, they find a lead, and then they go down that trail at that lead, then they find another lead, they go down that trail. And the way that this technology can effectively mimic that activity, as well as, in a way, kind of auto-summarize the output, is truly exciting for the compliance industry. But what we need to ensure is a couple of things.

0:32:08

One is the journalistic integrity, the trusted content that flows into these models. And then also, we take data privacy very, very seriously. And I think when organizations buy a product from Dow Jones, they know that we put it through its paces in relation to the various data privacy regulations. And that's truly important because you don't want to use a product that is potentially giving you misinformation or hallucinating or all the different pitfalls of AI, right?

0:32:39
Bernard Leong

Yeah, I think this is also a very good application of AI in a really important area that's actually affecting the world. You talk about just now your experience of speaking to the leadership in MAS. What would be your advice to, say, CISU executives or even government officials who are thinking about risk management with artificial intelligence?

0:32:59
Joel Lange

Well, I think it's just important to really understand kind of your technology providers, their view in relation to how they see themselves as a controller or a processor of information, really getting educated on the various data privacy laws and the interpretations around the world. We always say, particularly in relation to licensed journalism, don't commit a financial crime to combat financial crime, right? So don't go off and scrape data from the internet and take journalism and effectively use it to fight financial crime.

0:33:35

Because if you're scraping content, you are potentially violating intellectual property and copyright. And we're seeing with legislative developments in Australia, in Canada, a real respect for journalism in that regard. And that wave will continue. So I think it's just there are so many opportunities for efficiencies. But there is a duty of care there to understand that you're not following fall of other regulations when trying to comply with regulations.

0:34:06
Bernard Leong

So it comes to my traditional closing question then. What does CREIT look like for Dow Jones risk and compliance in the Asia-Pacific region, specifically with regards to the ASAM solution and also the market expansion?

0:34:19
Joel Lange

Well, we really want to continue with our focus on, in a way, uniquely understanding the regulations of individual countries and markets, having that kind of cultural specificity of our approach. And ultimately, organizations trust us. And they say that kind of the opposite of trust is verification. And with our brand, with our day in, day out, kind of focus on trusted journalism, as well as publishing content, which is going to help mitigate financial crime, be aligned to the regulations.

0:34:59

It's something that every day we need to prove that. And customers verified as well. We also have external audits. We have reviews of our content. But ultimately, it's about that trust. And so what we're really excited about with ASAM is that it's bringing together these core properties for the first time in a very efficient way, a federated search, the ability to do it at high volume. So it's bringing all those kind of key ingredients and core components and just making it much faster, much more efficient.

0:35:27

But keeping those underlying values, paying for content, paying royalties for content, building content to the regulations, you know, being on the balance of effectiveness and efficiency, having really distinct categories of content and filling those categories all the way to the top. I think that's something that really has set us apart and we're really excited to kind of continue to progress with.

0:35:51
Bernard Leong

That's a very thoughtful response, Joel. So many thanks for coming on the show and really have this really interesting conversation that covers actually very different areas from risk management, combating financial crimes, your ASAM solution, and also how artificial intelligence is being applied into this area of work. So in closing, I have two questions. So my first question is, any recommendations that have inspired your life recently?

0:36:17
Joel Lange

Well, that's a great question. You know, I love books and I recently read and it's going to sound like I'm trying to impress people in Dow Jones because it's actually a book that was written by the former editor of the Wall Street Journal. He's actually the editor at large now, Jerry Baker. But he actually has written a book recently that really hones in on kind of trust in institutions. And I think it really kind of encapsulates that last point around trust and verification, right?

0:36:49

And so what's so important for organizations right now is to be transparent in relation to the decisions they make and also, you know, live up to trust. And it's a tough time now in the world in terms of governments, in terms of the corporate community, in terms of various institutions about ensuring that that trust is there. And so I thought that was a really inspiring, you know, it's something we've been thinking about a lot. And I think that book, which recently came out, I very much recommend, you know, really hones in on that narrative.

0:37:22
Bernard Leong

There's a good recommendation. I'm probably going to check it out. So how can my audience find you?

0:37:27
Joel Lange

so i'm on linkedin i'm on i'm on twitter or now called x as well but we do a lot of linkedin activity that's probably the main main place but we do a lot of events around the world as well so definitely look out for us and we'd be great be great to interact with your community

0:37:43
Bernard Leong

definitely this video can be found on youtube and also on the audio will be available for all podcast platforms subscribe to us at Analyse Asia a-n -a-l -y-s -e asia in any platform of your choice, whether it's Threads, Mastodon, Tweet, or even Blue Sky. So Joao, many thanks for coming on to have this conversation with me, and I look forward to speak to you again soon. Thank you, Bernard.